Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
7This chapter presents research findings and, where applica- ble, product information relating to motor carrier risk avoid- ance strategies. Most of these are carrier efficiencies that also have benefits, or at least potential benefits, for safety. The chapter begins with a conceptual framework for Commercial Vehicle Operations (CVO) risk avoidance strategies based on the Haddon Matrix (discussed here) and on subsequent considerations of how that concept could be better fitted to motor carrier safety. The chapter then addresses the follow- ing carrier practices and issues: â¢ Employing preventive maintenance; â¢ Reducing empty (âdeadheadâ) trips; â¢ Minimizing loading, unloading, and related delays; â¢ Optimizing routing and navigation â Providing navigational and routing aids â Assigning familiar routes to drivers; â¢ Selecting road type: divided versus undivided roads; â¢ Avoiding work zones; â¢ Avoiding traffic; â¢ Emphasizing efficient scheduling: optimal times for safe travel; â¢ Avoiding adverse weather; â¢ Using higher-productivity vehicles; â¢ Using onboard computers and mobile communications; â¢ Employing team driving; â¢ Using EOBRs; â¢ Optimizing fuel economy and safety â Speed limiters â Monitoring driver fuel economy; and â¢ Monitoring vehicle condition. The last four topics in this list were not among the topics origi- nally planned for the study, but were added based on survey responses and interview inputs. The chapter ends with a dis- cussion of whether there is a general relationship between effi- ciencyand safety, in industryandinCMVtransport in particular. Three disclaimers are in order regarding the following discussion: â¢ No product or service was formally evaluated for this report. Company and brand names provided are illustra- tive of available products and services. Neither TRB nor this report endorses any company, product, or service. â¢ There are regulatory issues and activities under way regarding several of the earlier noted practices and equipment. This report addresses only their operational use by carriers, not regulatory questions. â¢ Project survey data are based on convenience samples of responding safety managers and other experts. Survey data represent the opinions and practices of the respon- dent samples, not of larger populations such as âall carrier safety managers.â Safety-manager respondents generally represented larger fleets with sufficient resources and safety interest to participate in national industry organi- zations and meetings. CONCEPTUAL FRAMEWORK FOR COMMERCIAL VEHICLE OPERATIONS RISK AVOIDANCE STRATEGIES This section provides a conceptual framework for carrier oper- ational risk avoidance strategies based on past literature. The Haddon Matrix (Haddon 1980) is a framework for under- standing and designing crash reduction strategies. It provides a conceptual structure for identifying factors that influence crash occurrence by dividing the crash scenario in terms of time frame (i.e., pre-crash, crash, and post-crash) and in terms of the primary âactorsâ affecting the event (Howarth et al. 2007). The âactors,â or categories of factors affecting crashes, are the human (primarily driver), the vehicle, and the roadway and environment. Thus, the conventional Haddon Matrix is a 3Ã3 matrix consisting of three rows (pre-crash, crash, and post-crash) and three columns (human, vehicle, and roadway and environment). Table 1 presents the conventional Haddon Matrix with examples in each cell. The Haddon Matrix was a seminal, heuristic contribution to motor vehicle safety and is a foundation for worldwide programs to reduce crashes (Williams 1999; Runge 2003). However, Will Murray (Murray et al. 2003, 2009) and others (e.g., Faulks and Irwin 2002) have pointed out that the conventional Haddon Matrix is insufficiently detailed for conceptualizing the full array of interventions applicable to CMV transport. Murray et al. (2003) added a column, âMan- agement Culture,â and listed 33 carrier, industry, and gov- ernment practices that affect safety, with most in the pre-crash time frame. Murray et al. (2009) reconceptualized and fur- ther expanded the columns to include six factors: â¢ Management culture; â¢ Journey; â¢ Road and site environment; CHAPTER TWO EVIDENCE AND PRODUCT REVIEW
â¢ People; â¢ Vehicle; and â¢ Society or community. Table 2 further reconceptualizes the matrix in the context of factors examined in the current report. As noted in the introduction to this report, âpre-crashâ encompasses several qualitatively different time frames: pre-trip, pre-threat, and pre-crash impact. The risk avoidance strategies addressed in this report are all either pre-trip or pre-threat, in that they are efficiencies and other practices that reduce the likelihood of imminent crash threats. The term âexposure avoidanceâ is used by one large trucking company contacted to character- ize these strategies. This is in contrast to pre-crash interven- 8 tions to prevent imminent crashes, such as forward collision warnings and similar crash avoidance systems. In Table 2, âpre-tripâ and âpre-threatâ are combined in one row because some practices fall across both categories. Another Haddon Matrix expansion in Table 2 is the sepa- ration of âpost-crashâ into âpost-crash/responseâ and âpost- crash/remediation.â Because most commercial drivers are company employees or representatives, the post-crash period extends to longer-term follow-ups, such as driver discipline and retraining. The current project surveys, interviews, and literature reviews make distinctions between crash risk factors that are TABLE 1 CONVENTIONAL HADDON MATRIX Time Frame Human (Driver) Vehicle Roadway/ Environment Pre-Crash Driver licensing Driver training, etc. Brake conditions Crash avoidance technologies, etc. Roadway markings Divided highways Curves, etc. Crash Restraint use Bone density, etc. Vehicle size Crashworthiness, etc. Guard rails Embankments, etc. Post-Crash Victim general health Rehabilitation, etc. Gas tank integrity Automatic collision notification, etc. EMS availability EMS response, etc. Ã¬Actor â/Factor Time Frame âActorâ/Factor Government/ Industry/ Society Motor Carrier Driver: Enduring Traits Driver: Temporary States Vehicle Design and Equipment Vehicle Condition Roadway Design/ Traffic Road/ Environment Condition Pre-Trip/ Pre-Threat Loading Delays Optimize Times of Travel Team Drivers Speed Limiters Monitor Fuel Economy Vehicle Size Onboard Computers & Comms. EOBRs Preventive Maintenance Monitor Vehicle Condition Deadheads Optimize Routing Divided Roads Exposure to Traffic Optimize Times of Travel Work Zones Exposure to Adverse Weather Pre-Crash Crash Post-Crash/ Response Post-Crash/ Remediation â â â â â â TABLE 2 CVO SAFETY MATRIX WITH CLASSIFICATION OF OPERATIONAL EFFICIENCIES AFFECTING SAFETY
9enduring (e.g., driver personality, vehicle design, and road- way design) and those that are temporary (e.g., driver rest status, vehicle condition, and weather). Accordingly, each of these three categories can be split into enduring versus tempo- rary, for finer classification of safety interventions. Additional columns are added for macro-level (government, industry, and society) and motor carrier factors, though most often these actors affect safety through specific effects on drivers, vehicles, or roadways. Table 2 includes these expanded breakouts and classifies the safety strategies of this report into an expanded CVO Safety Matrix based on the project review of the research and product literature, project surveys, and carrier interviews. All of the strategies addressed in this report are pre-trip and pre-threat interventions affecting one or more of the following factors: â¢ Driver temporary states; â¢ Vehicle design and equipment; â¢ Vehicle condition; â¢ Roadway design and traffic patterns; and â¢ Road environment condition (e.g., weather). In most cases, the safety benefits of the practices listed in Table 2 are well established. For at least two interventions (vehicle size changes and onboard computers and communi- cations), net safety benefits or disbenefits are not exclusively determined. They are still listed as strategies for considera- tion. One intervention, optimizing times of travel, relates strongly to two crash risk factors: driver temporary states (e.g., night driving during low circadian periods) and road- way design and traffic patterns (i.e., varying traffic density at different times). These factors may operate in opposite direc- tions at different travel times, thus complicating the problem of optimizing times of travel. PREVENTIVE MAINTENANCE Mechanical deficiencies are common in large trucks, reflec- tive of their large size, many components, and operational use. In the LTCCS, 40% of crash-involved trucks had some vehicle-related deficiency or malfunction, although these were the proximal cause (Critical Reason or CR) for only about 4% of crashes (excluding cargo shifts, which were another 2%). In the LTCCS, vehicle deficiencies as associ- ated factors were more common for combination-unit trucks (CTs) (43% of involvements) than for single-unit trucks (STs) (33%). There was a clear association with crash cate- gory and fault (CR assignment), as follows: â¢ Truck single-vehicle crash involvements: 62%; â¢ Truck multivehicle involvements, truck at-fault: 50%; and â¢ Truck multivehicle involvements, truck not-at-fault: 21%. These high percentages for vehicle condition as an asso- ciated factor were seen in the LTCCS because every crash- involved truck was given a full safety inspection. Other crash data files based on standard police investigations usually generate lower percentages because these investigations only note obvious system failures (Blower 2009). Roadside inspection statistics for FY2010 indicate that 19.5% of trucks and 6.5% of buses were placed out-of-service (OOS) owing to vehicle faults (FMCSA 2010). Note, however, that road- side inspections are targeted toward higher-risk carriers and thus do not represent a random sample of commercial vehi- cles in transport. In spite of these statistics, better carriers, including those accessible to this study through project surveys, generally have well-established and effective vehicle maintenance pro- grams. In well-managed fleets, as much as 80% of vehicle maintenance is planned rather than reactive (Arsenault 2010). Corsi and Barnard (2003) conducted a survey of âbest safety performersâ to identify and define their safety management programs and policies, including some practices covered in this report. They identified 148 safe motor carriers through a two-step process that included review of SafeStat perfor- mance data and obtaining recommendations from FMCSA State Division Directors. An extensive survey was completed by these 148 safe carriers and formed the basis for their report. The study found that 56% of these fleets used com- puterized equipment maintenance programs, with the per- centage ranging from 78% for the largest fleets to 23% for the smallest. Most (61%) of their computerized programs gener- ated specific part failure analyses. Such percentages would likely be higher today, given the advancement of technolo- gies and data systems supporting vehicle maintenance. Supportive attitudes toward fleet vehicle maintenance were strong in the Corsi and Barnard (2003) study. About 76% of carriers agreed or strongly agreed with the statement, âCost is no issue when it comes to keeping our vehicles defect-free.â About 80% agreed that, âDeploying a defect-free fleet is the most important thing we can do to ensure highway safety.â In CTBSSP Synthesis 1, Effective Commercial Truck and Bus Safety Management Techniques (Knipling et al. 2003), the project safety survey asked managers to rate the effectiveness of 28 fleet safety management practices. Regularly scheduled vehicle inspection and maintenance was rated the most effec- tive of the 28 practices. In the present project survey, 77 of 79 safety-manager respondents reported using a preventive main- tenance (PM) schedule and record for each vehicle, and 62 of 78 used PM software or spreadsheets. Both practices were rated among the most effective of the carrier practices pre- sented. Ironically, perhaps, vehicle condition was rated as among the factors with the least effect on overall crash risk among the five factors presented (enduring driver traits, tem- porary driver states, vehicle factors, roadway characteristics and traffic, and weather). This finding might partially reflect that, as with the two earlier surveys cited, the current survey drew its respondents primarily from among safety-conscious fleets. These results may characterize these better fleets, but should not be considered representative of the entire industry.
Greater involvement of drivers and other employees in vehi- cle management appears to have safety benefits for companies. Wright et al. (2005) analyzed safety programs in 12 Australian trucking companies. Firms that encouraged involvement of their workers in vehicle maintenance lowered maintenance costs, reduced crashes, and experienced less time spent by drivers away from work because of injuries. Four of the com- panies experienced reduced insurance costs and improved vehicle utilization through reduced OOS time. The project survey did not ask respondents to state whether the PM spreadsheets and other software they used were self- developed or commercially acquired. In the 11 case study interviews, several interviewees volunteered that they used commercially acquired maintenance management software. A review of these productsâ websites reveals numerous ways that truck maintenance software can assist fleets. These include helping fleets and other truck maintainers to better manage PM schedules and tasks, parts inventory, fuel and tire use, and other maintenance-related needs. These systems are marketed on the basis of reducing costs, improving productiv- ity, increasing warranty recoveries, improving auditing and billing, and generally making vehicle and other equipment maintenance more systematic. In most systems, information entered once is used in the system to support a number of dif- ferent user needs by populating various maintenance reports and schedules. For example, vehicle number is entered for every action on a vehicle, so that a complete vehicle mainte- nance history is always available. These records assist mainte- nance technicians in daily tasks and also support higher-level, fleet-wide analysis and planning. Specific data applications include: â¢ Scheduling PM; â¢ Generating work orders; â¢ Setting work standards for tasks; â¢ Tracking maintenance costs by vehicle, mile, hour, or other denominators; â¢ Recording equipment usage and inventory; â¢ Comparing equipment and maintenance procedure costs and reliability; â¢ Documenting licensing and inspections; â¢ Identifying and analyzing trends; â¢ Managing recalls; â¢ Purchasing parts and services; â¢ Bar-coding parts for further efficiencies; â¢ Managing warranties; â¢ Managing depreciation; â¢ Managing fuel use; and â¢ Tracking tools and other maintenance equipment. The product websites also include fleet maintenance improvement case studies and testimonials from maintenance managers. Among the current case studies, Carrier I provided the fullest explanation of their use of truck maintenance soft- ware. Carrier I has its own truck maintenance facility and man- 10 ages PM using TMT Fleet Maintenance software. The software is used to manage PM schedules, parts inventory, fuel and tire usage, and other maintenance schedules and records. Data on equipment assets and maintenance activities are entered once, and then integrated by the software into various user-formatted reports as an aid to equipment management and budgeting. In a Transport Topics editorial, Arsenault (2010) notes that vehicle maintenance is one of seven âBASICSâ in the new FMCSA Comprehensive Safety Analysis 2010 (CSA 2010) program. CSA enforcement practices increase the impor- tance of vehicle maintenance and the value of automating PM planning for carriers: As any fleet maintenance manager will attest, it is extremely dif- ficult to manually maintain a schedule of regular preventive maintenance services and inspections, track maintenance histo- ries, cross-reference driver complaints with repair orders and produce documentation on demand. The regulations donât man- date the use of maintenance management software, but it cer- tainly may make CSA 2010 compliance less arduous. . . . [The] software application simplifies this process and makes it much easier to comply with CSA 2010. Question 30 of the safety-manager survey asked respon- dents to identify the operational efficiency or other practice contributing most to fleet safety. The following are responses relating to the general topics of vehicle maintenance and inspections: â¢ Conducting a thorough pre- and post-trip inspection; â¢ Proper pre- and post-trip inspections, driver debriefing and communication; â¢ Insistence on daily management monitoring of pre- and post-trip inspections; â¢ Driver and maintenance staff input; and â¢ Preventive maintenance and pre- and post-trip inspec- tions. REDUCING EMPTY (âDEADHEADâ) TRIPS One of the simplest ways to improve safety through improved efficiency is to reduce the number of unproductive, non- revenue trips and miles. Reducing empty miles is primarily motivated by financial gains, but there is a proportional benefit to safety. The textbox provides a simple hypothetical model to illustrate how safety is enhanced by reducing empty backhauls. Reducing empty trips is one of the operational practices being examined in the Motor Carrier Efficiency Study (MCES). The MCES (Delcan Corporation 2007a,b; Belella et al. 2009) is a congressionally mandated program to identify inefficiencies in freight transportation, evaluate safety and pro- ductivity improvements made possible through wireless tech- nologies, and demonstrate wireless technologies in field tests. Phase I of the study has gathered extensive data on carrier inef- ficiencies in seven categories: equipment and asset utilization, fuel economy and waste, loss and theft, safety (i.e., crashes),
11 maintenance, data and information processing, and business (including driver) management. Phase II will involve pilot tests of specific efficiency improvements. Reducing empty miles falls under the MCES category of equipment and asset utilization. The project literature review report (Delcan Cor- poration 2007a) states the following rationale: Motor carrier equipment is effectively utilized when it is in the process of generating revenue, by the mile, the hour, or by any other mutual agreement between the motor carrier and their cus- tomer. Ideally, equipment would operate around the clock with neatly planned and minimal down time for routine maintenance, repairs, or refueling; equipment on-the-clock for the purposes of serving the customer would always be compensated, there would be no deadhead, unauthorized, or out-of-route miles, and trucks would never have to wait for or travel empty to pick up the next load. However, asset utilization is not optimized in real world operations. Additionally, disconnects between certain types of motor carrier operations and shipper/receiver/customer opera- tions exist as continuing impediments to optimal asset utilization. The MCES (Delcan Corporation 2007a) quotes several estimates placing empty miles at about 20% of total miles for for-hire carriers. The percentage appears to be decreasing slowly as companies do a better job of obtaining backhauls. The report cites a National Private Truck Council (NPTC) estimate that the percentage is about 25% for private fleets. Drayage operations have higher percentages, with empty trucks adding considerably to traffic congestion in the vicin- ity of major U.S. ports. The report estimates total U.S. large- truck empty miles at 40 billion annually. It estimates the potential financial gain to carriers from eliminating empty miles to be $2.7 billion. In 2008, the U.S. large-truck fatal crash rate was 1.64 per 100M vehicle-miles traveled (VMT), and the fatality rate was 1.86 per 100M VMT (FMCSA Analysis Division 2010). This suggests that approximately 656 fatal crashes and 744 fatalities that year involved empty trucks. The total truck crash rate was 160.4 per 100M VMT. Applying this rate to 40 billion empty miles suggests approximately 64,000 asso- ciated crashes. These estimates, like the textbox model pre- sented previously, assume that crash rates are unchanged when trucks are empty. For-hire carrier sales and dispatching focus in large part on reducing empty miles. Delivery contracts are often written to establish routes that minimize empty miles. Private fleets may function as for-hire fleets by seeking backhaul loads. Carriers obtain backhaul loads through load brokers, load boards, and development of long-term service contracts with shippers. Economies of scale favor larger carriers in aggressively reduc- ing empty miles, with some large firms attaining empty ratios of 10% (Delcan Corporation 2007a). Reducing empty back- haul rates is also popular with drivers because, in most oper- ations, they are not paid for empty miles. Web-based load boards (also called freight boards) and load brokers are an efficient and economical means for match- ing loads to trucks. Service providers offer round-the-clock online service, route searches, shipment tracking, routing aids, credit reports, and both carrier and shipper quality rat- ings based on user feedback. Uship, for example, provides a profile and customer feedback-based ratings for each carrier (by shippers) and for each shipper (by carriers). Individual comments on carriers from shippers, and on shippers from carriers, are listed. For each carrier, there is an overall feedback score, positive feedback percentage, and customer ratings on four scales (communications, care of goods, punctuality, and service as described). In the project surveys, reducing empty backhauls was regarded by safety managers as having modest benefits for fleet safety. Respondents were asked to rate the safety bene- fits of reducing empty miles on a seven-point Likert scale ranging from â3 (âReduces Fleet Safetyâ) to +3 (âImproves Fleet Safetyâ). The overall safety-manager mean rating was +0.5. A later question asked whether the carrier, âUse[d] bro- kers of other services to reduce empty backhauls (dead- heads).â Among truckload carrier respondents, 21 of 27 indicated that they did. Among users of such services, the prac- tice received a 3.1 mean Likert scale rating on a 1-to-5 scale of safety effectiveness. Most for-hire carriers represented in the project case stud- ies made strong efforts to minimize their empty miles. One large company, Carrier B, has taken advantage of its size and used multiple means to a 10% deadhead rate, which the inter- viewee considered to be a major accomplishment. Carrier E, a medium-sized truckload carrier, has a 12% deadhead rate, which has allowed it to pay its drivers for empty miles. This Empty Backhauls and Safety: A Hypothetical Model A simple, âback-of-the-envelopeâ math model illustrates the safety gains, relative to productivity, from reducing dead- head trips. Suppose, hypothetically, that a trucking fleetâs miles are 80% full and 20% empty. Assume that all its miles (full and empty) carry the same crash risk. During Year 1, the company has 120 police-reported crashes carrying 8 million ton-miles of freight. In relation to productivity, that equals a rate of 15 crashes per million ton-miles. During Year 2, it uses the same trucks and same drivers, and logs the same number of total miles. It makes no safety improvements in its individual trucks and drivers, and has the same crash rate per mile. Thus, it again has 120 crashes. But it uses load brokers, load boards, and other means to reduce its empty miles by half. Thereby, the company runs 90% full and 10% empty, and carries 9 million ton-miles of freight. During Year 2, its crash rate in relation to productivity is 120 crashes per 9 million ton-miles or 13.3 crashes per ton-mile. That is an 11% improvement from Year 1. These hypothetical calculations are described for one fleet, but in principle extrapolate to the entire national fleet. They illustrate the safety effects of reducing deadheads and, more generally, how carrier efficiencies can result in risk avoidance apart from traditional risk-reduction efforts.
eliminates one source of driver discontent and improves driver retention. Improved driver retention, though outside the scope of this report, is itself an operational efficiency that is highly supportive of safety (Staplin et al. 2002; Knipling 2009). MINIMIZING LOADING, UNLOADING, AND RELATED DELAYS In almost any work activity, waiting is an unproductive use of time and reduces efficiency. Excessive delays associated with truck loading and unloading, also known as driver deten- tion, affect safety as well. Drivers generally are unable to use wait times for sleep or other restorative rest. Thus, wait times use up driversâ available waking hours, thereby contributing to later fatigue when they are finally driving. Further, drivers may be thrown off-schedule by excessive waits, thus caus- ing frustration and a later urge to drive faster or otherwise increase work speed unwisely. Under current hours-of-service (HOS) rules, driversâ tours-of-duty are limited to 14 hours. This has the benefit of preventing longer work periods, but raises the potential for drivers to rush to reach a destination within 14 hours. A new report by the U.S. Government Accountability Office (GAO 2011) addresses the issue of commercial driver detention times. GAOâs summary findings included: â¢ Detention of drivers at shipper or receiver facilities is a prevalent problem; of 302 drivers interviewed by GAO, 204 (68%) reported being detained within the past month. â¢ Of those drivers who had experienced detention, 80% stated that it affected their ability to meet HOS require- ments, and 65% reported losing revenue as a result of being detained. â¢ Shippers and receivers control many of the factors lead- ing to driver detention, such as facility staffing, loading and unloading equipment, quality and promptness of service, and the readiness of products for pickup. â¢ Shippers often disagree with carriers and drivers about the length of detention time and its causes. â¢ Carriers have some ability to mitigate the problem by charging detention fees to shippers, developing better working relationships with customers, improving com- munications, and abandoning shipper accounts where detention is a problem. â¢ The âhook and dropâ method, whereby a truck arrives with an empty trailer and leaves with an already-loaded trailer greatly reduces the problem, but requires more equipment, coordination, and space. â¢ Larger carriers have greater resources and more lever- age with clients than smaller carriers, and thus are gen- erally able to mitigate the problem more effectively. â¢ The quantitative contribution of driver detention to HOS violations and to crashes is not known. Under the sponsorship of the FMCSA, the Trucking Research Institute conducted an experimental study of the 12 effects on driving alertness of truck loading and unloading tasks (OâNeil et al. 1999). There was no consistent evidence of driving fatigue resulting from the physical activity. Instead, drivers complained about the time required and unplanned delays associated with loading and unloading. Moreover, because drivers in many segments of CMV transport do not load and unload their vehicles, the question of excessive phys- ical work is often moot. Instead, the problem revolves around detention times. CTBSSP Synthesis 1: Effective Commercial Truck and Bus Safety Management Techniques (Knipling et al. 2003) asked safety-manager and other-expert respondents to rate the relative importance of 20 CMV transport safety prob- lems. The problem was stated as, âDelays associated with loading and unloading (e.g., resulting in long working hours, tight schedules, and fatigue).â In Likert scale ratings, the item was judged the fifth most important safety problem by safety managers and the fourth most important by other experts. About half of the respondents considered it one of the top five problems among the 20 presented. The MCES Inefficiencies Report (Delcan Corporation 2007b) identified operational inefficiencies recognized as most pressing by motor carriers, cited evidence of their effects, and evaluated potential technological solutions. These ineffi- ciencies were defined as practices, procedures, incidents, or events that produce waste, generate unnecessary expenses, require excess effort, do not generate revenue, or do not con- tribute to safe, secure, and timely cargo transport. The MCES study team conducted stakeholder workshops in seven U.S. locations in which representatives from motor carriers, tech- nology vendors, and other industry experts discussed transport inefficiencies. Excessive waiting for loading and unloading was the most frequently cited âhigh-priorityâ inefficiency across the stakeholder groups. This was the top inefficiency concern of truckload (TL), less-than truckload (LTL), and intermodal carriers (Belella et al. 2009). Carriers expressed particular frustration regarding excessive waits for their trucks to be unloaded at consignee locations as well as at intermodal terminals. Border crossing wait times were also cited. Loading and unloading inefficiency is costly for carriers and their drivers, who routinely bear the expense of waiting. Delcan Corporation (2007b) estimated the average truck wait- ing, loading, and unloading time at pickup and delivery points to be 2 hours, with much of the time spent waiting. According to the report, there is a potential annual financial gain of $3.1 billion for U.S. carriers and $6.6 billion for society as a whole from the elimination of this transport inefficiency. The most affected CMV transport operations are motor carriers of containers (e.g., port drayage operators), regional and long- haul TL carriers, and grocery and agricultural LTL carriers. For deliveries, the problem often affects private carriers as much as it does for-hire carriers, because most of their deliv- eries are to customers whose trucks are treated the same as for-hire trucks. The problem is less significant for private car-
13 riers loading at their own facilities and for those delivering to their own stores or other facilities (Delcan Corporation 2007b). According to carriers reporting to the MCES and those interviewed for the current study, shippers and receivers may be relatively indifferent to the costs incurred by carriers and drivers while waiting. Further, shippersâ and receiversâ own efficiency may actually benefit from practices that create steady truck queues while their own operations proceed on schedule without workload spikes or interruptions. Gate reser- vation and appointment systems could alleviate the problem, although carriers reporting to the MCES argued that their use was geared toward optimizing facility efficiency rather than reducing truck waiting times. One corporate vice president for safety interviewed for the study believed that carrier use of EOBRs had an indirect benefit of reducing loading and unloading delays. That is because EOBRs reinforce the notion that HOS compliance is nonnegotiable, and also because they provide more com- pelling documentation of delivery schedule disruptions caused by excessive dock delays. The MCES (Delcan Corporation 2007b) suggested a wire- less communications application concept called Virtual Queu- ing as a logistical and technological intervention to reduce excessive loading and unloading delays. Virtual queuing would extend queues to trucks reaching the vicinity of the terminal. It would allow consignees to monitor and dynam- ically reschedule dock operations to compensate for delays in both truck arrivals and departures from the facility. On project surveys, respondents were asked to rate the safety benefits of âreducing loading and unloading delaysâ on a seven-point Likert scale ranging from â3 (âReduces Fleet Safetyâ) to +3 (âImproves Fleet Safetyâ). Among safety managers, the overall mean rating was +1.8. Among other- expert respondents, the mean rating was +1.4. For both groups, these were among the highest mean ratings for driving situa- tions and practices presented. Several of the case study carriers assign dedicated routes to successful, experienced drivers. These assignments are coveted. In addition to keeping drivers on familiar roads, dedicated runs provide a stable income, predictable home times, and more regular work-rest cycles. For carriers, the most obvious and feasible means to reduce excessive delays is to charge a fee to shippers and receivers for excessive wait times. These detention fees are written into shipping contracts, with 2 hours appearing to be the most common threshold for fees. Affected drivers may receive all or most of the money charged. Safety-manager respondents were asked whether they charged detention fees to customers for excessive loading and unloading delays. The question was omitted from bus operator forms. Among all responding carriers, 34 of 50 charged detention fees. Among TL carrier respondents, the proportion was 23 of 27. The safety effec- tiveness of the practice was given a mean rating by users of 3.4 on a 1-to-5 Likert scale. Several interviewees for case studies used detention fees. One interviewee considered them effective, but added that âaggressiveâ enforcement and col- lection was essential for reducing excessive delays and their negative safety consequences. Another lamented that the practice was not highly effective because the fees became a corporate-to-corporate billing issue, rather than a penalty felt directly by frontline depot supervisors with the most influence on the problem. The following are two safety-manager survey comments relating to loading and unloading delays: â¢ Inefficiencies of shippers/receivers in the loading and unloading process has the most negative effect on safety for our drivers. â¢ Our biggest challenge is with our customers and suppli- ers. There is an ignorance or apathy towards an efficient loading or unloading process. OPTIMIZING ROUTING AND NAVIGATION Smoother routing and navigation improve the efficiency of CMV operations. The following quotation from the MCES literature review report (Delcan Corporation 2007a) summa- rizes the importance of optimized routing for operational efficiency: âEach time a truck accrues additional miles due to less than optimal routing, the equipment is not being utilized to its full potential, and it does not complete its intended mis- sion in the least possible time, with the least possible costs in labor, equipment wear, and fuel.â Improved routing effi- ciency appears to aid safety as well. There are two primary safety rationales for aiding routing and navigation. The first is avoiding exposure, especially to higher-risk roads, that is, roads likely to be congested with traffic or otherwise haz- ardous. The second is easing driversâ navigation workloads. Making navigation easier for drivers reduces distraction and associated crash risk. This section explores these benefits and describes products aiding CMV routing and navigation. A distinction can be made between routing and naviga- tion in CMV operations (Bennett 2009). Routing optimization generally refers to improvements in the efficiency of an over- all pickup-and-delivery sequence, as in a full driver tour- of-duty or multiday trip. Routing optimization can also be applied to a whole fleet or company. Navigation aids more often refer to devices to assist drivers in making a particular point-A-to-point-B trip. Providing Routing and Navigational Aids to Drivers Portable and vehicle-installed Global Position System (GPS) devices are marketed as aids to navigation and mobility, not
as safety devices. Yet the proper use of automated GPS nav- igation aids by commercial drivers supports safety by the two mechanisms mentioned earlier: reduction of risk exposure and easing driver mental workload. With regard to risk expo- sure, reductions can simply be in the quantity of exposure (i.e., reducing mileage for the same productivity) and in exposure to higher-risk road conditions. Truck-specific navigation aids can steer drivers clear of roads where truck traffic or hazardous cargo is restricted or prohibited. They can warn drivers of low-clearance under- passes (e.g., bridges with less than 14 ft of vertical clearance, the national standard for local roads and collectors), low- weight-bearing bridges, or other hazardous locations. Sys- tems can route drivers around higher-risk roads, such as undivided roads and those with high traffic densities. If sys- tems are updated, they can route drivers around work zones or road closures. The relative risks associated with some of these road types and conditions will be documented in the following sections. Systems can also route trucks to avoid toll roads, although this diversion practice is more likely to be detrimental to safety, as toll roads generally have safer design features than do non-toll alternate routes (Short 2006). Any system that routes trucks away from higher-risk roads and toward lower-risk roads reduces overall crash risk inde- pendently of driver and vehicle risk factors. Truck-specific road information is needed, however. Leone (2010) reports the experience of trucking company Transport America, two of whose trucks were involved in bridge underride crashes in the preceding year. In one case, the driver was following a paper map, and in the other, the driver was following a general-purpose navigation device. Neither driver was aware that the route included low-clearance underpasses. The safety manager interviewed for Case Study D (Large Truckload Carrier) told of similar mishaps related to truck drivers using general-purpose navigation aids. A bill proposed (though not under immediate consideration) in New York state would out- law commercial drivers from using general-purpose GPS units (Leone 2010). Multiple vendors provide truck-specific routing and navi- gation aids. Major communications providers can offer inte- grated, truck-specific navigation systems with their systems. Specific features and services of truck-specific routing and navigation aids may include: â¢ U.S.- and Canada-wide street-level map data with turn- by-turn directions; â¢ Routing in accordance with height restrictions, low- weight bridges, seasonal road closures, and so forth; â¢ Best âpracticalâ versus shortest routing choices; â¢ Optimized stop sequences; â¢ Truck-specific toll costs; â¢ Fuel optimization for cost and in accordance with com- pany purchase plans; â¢ Hazardous materials and larger-truck-size routing; 14 â¢ Intermodal rail versus truck routing and cost compar- isons; â¢ Customized routing by specification of preferences or routes to be avoided; â¢ Manual override of specified routing; â¢ âLocation radiusâ tool to search for points of interest within specified distances of any location; â¢ Geofencing (restriction of vehicles within or outside of specified zones); â¢ Driver rest stop options in compliance with HOS rules; â¢ Identification of weigh stations along routes; â¢ Identification of intermodal rail ramps; â¢ Automation of fuel and mileage tax tracking; â¢ Integration with communications systems for real-time management; â¢ Web access to routing and navigation functionality; â¢ Data downloads to spreadsheets or other programs; and â¢ (Being developed) Dynamic adjustments based on real- time assessments of traffic and weather conditions. Potential GPS benefits touted by vendors include better real-time dispatching, increased productivity, improved HOS compliance, decreased overtime, lower fuel use, improved customer service, validation-of-service calls, lower insurance costs, and decreased driver speeding. One deficiency of most truck-specific navigation systems is that they route to an address, not to a delivery entrance. For large delivery loca- tions, this means that drivers may still be at a loss to pinpoint their exact destination. Manual âlast-mileâ directions are still needed (Leone 2010). Another, more significant concern is that different vendors may gather their own highway data themselves or through independent contacts with state and local agencies. There is little standardization across different vendors and products. In an article on route optimization benefits, Bennett (2009) quotes Ken Snow, president of Hagopian Cleaning Services, on the companyâs successful use of route optimization soft- ware. The companyâs fleet consists of 27 vans used for carpet and upholstery cleaning. Snow estimates that route optimiza- tion reduces company mileage by 5% to 10%, or 25,000 to 50,000 annual miles. This is reduced exposure to crash risk. Route optimization also provided fuel savings and the elimi- nation of the manual task of route planning for the trucks. Each evening, the company runs the route optimization program to plan the next dayâs routes. Onboard GPS units would aid drivers in A-to-B trips, but would not optimize the sequences of multiple stops of a truck or multiple trucks of a fleet. Larger companies with multiple fleet locations can network the appli- cation to optimize regional coverage. A vendor interviewed for the article claimed that companies with as few as four vehicles could benefit from route optimization software. For mobile maintenance or other service operations, trips back to a central depot can be reviewed to determine whether they could have been avoided by better provisioning of the vehicle before its departure. Another company representative interviewed for the article reported the results of a company survey of drivers
15 using a routing and navigation aid system. The survey found that 85% of company drivers believed the system helped them every day to travel more efficiently and with less stress. On project surveys, respondents were asked to rate the safety benefits of using routing and navigation aids on a seven-point Likert scale ranging from â3 (âReduces Fleet Safetyâ) to +3 (âImproves Fleet Safetyâ). The item was stated as, âIncrease routing efficiency using GPS navigation aids and/or truck routing software.â Among safety managers, only one respondent out of 77 assigned the practice a nega- tive rating, and the overall mean rating was +1.8. Among other experts, there were no negative ratings among 31 respondents and the mean rating was +1.1. With regard to carrier practices, 42 of 76 respondents said their drivers use general-purpose GPS systems, and 29 of 77 used truck-specific systems. There was considerable overlap among the users of the two types, suggesting that in many fleets some drivers use general systems, whereas others use truck-specific systems. Assigning Familiar Routes to Drivers Driving involves three types or levels of performance and skill: controlling the vehicle, responding to driving events (e.g., other traffic, signs, and signals), and navigation. Exces- sive attention to any one of these levels can interfere with performance on the others. For example, novice drivers are typically preoccupied with controlling the vehicle, to the detri- ment of their ability to respond to traffic events (Shinar 2007). Similarly, excessive attention to navigation can reduce atten- tion to basic vehicle control and, in particular, traffic events. Think about your own driving on roads that you drive every day, compared with your driving on unfamiliar roads. For experienced drivers on familiar roads, both vehicle control and navigation are automatic. One can anticipate and attend closely to traffic conditions and specific threats. On unfamiliar roads, drivers are not as able to anticipate specific roadway and traf- fic risks. This is particularly true when drivers are looking for a specific turn or destination. Situation awareness is the ability to âeffectively filter information in a data-rich environment,â or simply âknowing what is going onâ (Shinar 2007). The primary task of driving includes controlling the vehicle and responding to traffic events. Any secondary task such as navigation can reduce performance on the primary task. LTCCS statistics showed a relation between truck driversâ roadway familiarity and crash involvement. The LTCCS had no mileage exposure database or non-crash control data set, so relative rates of crash involvement cannot be discerned. One can, however, use LTCCS data to discern associations between roadway familiarity and fault in crashes, with the assumption that a higher incidence of fault also implies higher risk. Figure 3 shows three categories of LTCCS truck crash involvements and, for each, the percent of truck drivers who were unfamiliar with the road. That is, they had never or only rarely driven the road before. Overall, 26% of LTCCS drivers were unfamiliar with the roads on which they crashed. When crash involvements were disaggregated by driver unfamil- iarity with the roadway, one sees a strong relationship with crash fault and type of crash involvement. On project surveys, respondents were asked to rate the safety benefits of assigning familiar routes to drivers when possible. As with similar questions, respondents were pre- sented with a seven-point Likert scale ranging from â3 (âReduces Fleet Safetyâ) to +3 (âImproves Fleet Safetyâ). The overall mean for safety manager-respondents was +1.7. Among other experts, the mean rating was +1.6. Beyond roadway familiarity, are there other strategies managers and dispatchers could use in assigning routes to drivers? One prudent strategy is to assign more difficult and risky routes to more experienced and competent drivers. Or, stated another way, avoid the situation in which an inexperi- enced or otherwise questionable driver is exposed to higher- risk traffic or roadway conditions. Knipling (2009) presented a hypothetical mathematical model suggesting that this strat- egy would reduce overall fleet crash risk because it would reduce the dangerous convergence of the weakest drivers with the most hazardous roadway situations. The model assumed a 38% 29% 17% 10% 15% 20% 25% 30% 35% 40% % R ar el y or N ev er Dr ov e Ro ad 0% 5% Single-Vehicle Crash MV Truck At-Fault MV Truck Not At-Fault Truck Crash Involvement Category FIGURE 3 Road unfamiliarity and crash involvement.
multiplicative relation between relative driver risk and rela- tive road risk. ROAD SELECTION: DIVIDED VERSUS UNDIVIDED ROADS Recall the speed paradox presented and discussed in chapter one. Statistics were presented showing that, perhaps contrary to intuition and expectations, truck travel at speeds above 50 mph was generally far safer than travel under 50 mph. A principal reason underlying the speed paradox is that most higher-speed truck travel is on divided highways, whereas most lower-speed travel is on undivided roads. Further, most divided highways have limited access (i.e., entrance- and exit-ramps), whereas undivided highways are open-access. The safety advantages of divided over undivided highways are well known to road designers and others in road safety. About 85% of large-truck crashes involve another vehicle, and interaction among vehicles is greatest on undivided roadways. On undivided roads there are traffic signs and signals, crossing traffic, stops and starts, turns, pedestrians and bicyclists, and generally greater opportunities for distractions and other driver mistakes. On divided highways, vehicles are all traveling at about the same speed with minimal interaction and few crossing-path events. Thus, divided highways have much lower crash rates. Overall, Interstate highway fatal crash rates are about one-half those of non-Interstate arterial roads, and just one-third those of local roads (FHWA 2000). In eight studies cited in the FHWA website on crash modification fac- tors (www.cmfclearinghouse.org), the road design counter- measure âinstall roadway medianâ reduced crash rates by an average of 49%. Harwood (2006) emphasized the limited-access feature of most divided highways in the following comments regarding roadway design and CMV safety: The lowest crash rates on our roadway systems are on limited- access highways, for example, freeways and toll roads. Higher crash rates occur on multilane non-freeways where direct access is permitted, including both multilane divided highways and multilane undivided highways. Two-lane highways have the highest crash rates. Across the mix of highway types, crash rates differ by at least a factor of 3 or 4 between typical rural two-lane highways and rural freeways. Naturalistic driving data provide a compelling testimony to the risks of driving on undivided versus divided roads. Traffic conflicts (including crashes, near-crashes, and other incidents) captured by onboard sensors and videos are classified by their conditions of occurrence and other characteristics. In this case, the condition of interest is roadway separation (divided versus undivided). In addition, researchers randomly select a large sample of exposure observations, or âexposure points,â repre- senting normal driving. These âexposure pointsâ are also clas- sified by conditions of occurrence. Researchers then compare the frequencies of conditions (e.g., divided versus undivided 16 highways) in the traffic conflict sample with those of the expo- sure points. Any condition overrepresented in the traffic con- flict sample can be considered a safety risk factor. These comparisons demonstrate the disproportionate risk associated with driving on undivided highways. Table 3 com- pares 1,072 baseline epochs (representing exposure) to 907 traffic conflicts (crashes, near-crashes, and other incidents) from a long-haul truck naturalistic driving study (Hickman et al. 2005; Knipling et al. 2005). The percentage breakdowns are shown. Most divided roads in the study were Interstates or other freeways. Only 10% of the driving was on undivided highways, but 38% of the traffic conflicts occurred on those roads. The majority of traffic conflicts occurred on divided highways (62%), but the risk relative to exposure was much greater on undivided roads. The odds ratio between undivided and divided roads for traffic conflicts was calculated to be 5.3 [(248/113) / (559/959) â 5.3]. Figure 4 shows the same relationship graphically. In the LTCCS, 38% of CT crashes occurred on undivided (including one-way) roads, a percentage identical to that in the naturalistic driving study. Percentages were similar for single-vehicle and multivehicle CT crashes. A larger percentage of LTCCS ST crashes (55%) was on undivided roads. STs have somewhat higher crash rates per mile trav- eled than do CTs, reflective of their greater exposure on undi- vided roads. In 2008, 54% of large-truck fatal crashes and 52% of nonfatal crashes occurred on undivided highways (FMCSA Analysis Division 2010). No national exposure estimates using the same definitions are available for com- parison to crashes, but there is no question that the exposure percentage is much smaller. These statistics indicate that carriers could reduce risks by making concerted efforts to dispatch and route trucks on divided highways. For load pickups and deliveries, it may not be possible to reduce exposure significantly to undi- vided highways. Greater opportunities exist in trip planning between loading and delivery points. The statistics suggest that, when given a choice, trucks are safer on divided high- ways even if that means significantly greater mileage. Event Type Roadway Type Exposure (%) (N = 1,072) Traffic Conflict (%) (N = 907) Undivided 113 (10) 248 (38) Divided 959 (90) 559 (62) Total 1,072 (100) 997 (100) TABLE 3 UNDIVIDED VERSUS DIVIDED HIGHWAYS: COMPARISON OF NATURALISTIC DRIVING EXPOSURE POINTS TO TRAFFIC CONFLICT DATA
17 An extension of the earlier findings and logic applies to the toll road avoidance strategy known as âdiversion.â Diver- sion (Short 2006) occurs when trucks or other vehicles eschew toll roads to avoid paying the tolls. Often this is on a parallel undivided highway. There are about 2,800 mi of toll roads within the 42,800-mi Interstate highway system, and about 1,800 additional mi of non-Interstate toll roads (Short 2006). Many toll roads operate at below-capacity volumes, in part as a result of diversion. No one knows how many trucks and other vehicles divert from these roads to avoid tolls, or the effects of diversion on overall crash rates. It is important that carriers and drivers carefully weigh their road choices by factoring relative crash risks into their decisions. In the safety-manager surveys, âmaximizing travel on Inter- states and other freewaysâ was among four factors tied for the highest mean rating (+1.8 on a â3.0-to-+3.0 scale) of the 11 driving situations and practices presented. Its opposite, maximizing travel on low-speed roads, received the lowest mean rating (â1.1). For other experts, the corresponding val- ues were +1.7 (the second highest rating for 15 practices pre- sented) and â1.6 (the highest negative rating). With regard to toll reimbursements to drivers, 66 of 78 responding carrier rep- resentatives indicated that they provided âEZ Pass transpon- ders and/or reimbursement of toll charges to drivers/OOs [owner-operators].â The practice received an average rating of 3.8 on the five-point safety effectiveness scale. On the other-expert survey, a respondent provided the following comment (with no sources cited for statistical statements): The biggest single determinant of overall safety is risk exposure. Interstates, because they are divided trafficways with no at-grade intersections, are 400% safer than U.S. and State routes. More than 70% of fatal truck crashes occur on these latter roads, not Interstates where all the enforcement attention and focus takes place. Carriers operating mostly on non-Interstate roads are much more at-risk than those that predominantly travel up and down Interstates. Almost all of the case study carriers equip their vehicles with toll transponders and actively encourage travel on divided roads. Several use routing software maximizing travel on these roads. Others pre-plan trips in detail, including road choices. AVOIDING WORK ZONES In 2008 there were about 10,000 large-truck crashes in work zones, about 3% of all truck crashes. These included 166 fatal crashes, 4.4% of the total of 3,733 for the year (FMCSA Analysis Division 2010). About one-fourth of all work zone fatal crashes involved a large truck. Roadway- and traffic- related crash threats in work zones include constricted lanes, narrow or absent shoulders, makeshift signs, and traffic back- ups where light vehicles may dart in front of trucks to move up in the queue. An FHWA website (www.workzonesafety.org) contains crash data, research reports, driver training materials, and other information on work zone safety, including infor- mation on major highway work projects. In the same naturalistic driving data as discussed earlier (Hickman et al. 2005), trucks drove in highway work zones in only 8 of 1,072 randomly selected exposure-point obser- vations (0.7%). During the same driving, they had 55 of 914 traffic conflicts in highway work zones (6.0%). Although these data are based on relatively few work-zone observa- tions, they suggest greatly elevated risk. The calculated odds ratio for conflict involvement in this data set is 8.5 [(55/8)/(859/1,064) â 8.5]. The elevated crash risk in work zones is not specific to trucks. Khattak et al. (1999) found that changes in crash rates during the construction were about the same for cars and trucks. Truck crash severity was not increased by roadwork, but work zone crashes tended to involve more vehicles than those on normal roads. 0.9 0.1 0.38 Undivided Roads Divided Highways 0.62 Exposure Incidents FIGURE 4 Undivided and divided roads.
Murray et al. (2005) analyzed work zone crashes and sug- gested truck-related improvements. The study found rear- end and sideswipe crashes to be among the most common scenarios. Work zone fatal crashes are more likely than non- work zone crashes to involve multiple vehicles. Nearly one- third of fatal work zone crashes involve a truck, although the study did not determine relative fault or principal causes. Work zone crash countermeasures suggested by the study included crossroad rumble strips, driver feedback signing (warning of excessive speed), highway advisory radio, and detection and warning of traffic queues. Thirteen percent of truck-crash involvements in the LTCCS occurred in work zones. Almost all of these involvements were in multivehicle crashes. Trucks were assigned the CR (were at-fault) in 42% of these. Of all truck at-fault LTCCS crashes, 11% occurred in work zones. Many of these were rear-end crashes in which trucks struck cars, suggesting lia- bility for trucks and their carriers. In the safety-manager survey, avoiding construction zones received an average rating of +1.4 on the â3-to-+3 Likert scale. For other experts, the mean rating was +1.3. In both cases, it was in the top half of safety practices but not among the very top. In safety-manager interviews, work zones were cited several times as being among the risky road conditions to be avoided for safer operations. Two carriers described specific efforts to avoid them. One carrier codes work zones on its internal crash reports and has identified them as high- risk areas. Another carrier provides drivers with daily state traffic alerts that include information on major work zones. AVOIDING TRAFFIC The speed paradox (chapter one) and other evidence pre- sented previously suggest that disrupted traffic flow elevates crash risk. Heavy traffic has become a dominant feature of urban travel. The Texas Transportation Institute (TTI; http:// mobility.tamu.edu) publishes annual reports on urban traffic congestion and its effects on mobility (Schrank and Lomax 2009). In 2007, Americans lost 4.2 billion hours to urban con- gestion. This was a small reduction of about 1% from the pre- vious year, but was still more than five times the urban delay 25 years ago. Across the United States, delay has increased in all types of urban areas, whether relatively small, medium- sized, or large. In larger urban areas, free traffic flow occurs reliably only between the hours of 9:00 p.m. and 5:00 a.m. In 1982, peak morning congestion lasted about 75 min, from about 7:30 a.m. to 8:45 a.m. Equivalent congestion now lasts almost 3 hours from about 6:30 a.m. to 9:15 a.m. For evening peak hours, 1982âs 90-min peak, between about 4:00 p.m. and 5:30 p.m., is now seen for twice as long, between about 3:30 p.m. and 6:30 p.m. The Travel Time Index is the ratio of travel time in the peak period to travel time at free-flow conditions. Since 1982, the index has risen steadily from 1.09 to 1.25. That means that urban travel times during peak hours are 25% slower than during free-flowing conditions. 18 Increases in traffic density and travel times generate disproportionate increases in the number of proximal inter- actions among vehicles and associated crash risk. This is perhaps best seen in naturalistic driving data. Large-truck nat- uralistic driving methodologies and statistical findings relat- ing to traffic density and risk are similar to those presented earlier for undivided highways and for work zones. Table 4 shows exposure and traffic conflict percentages for different levels of traffic density from Hickman et al. (2005). As with earlier examples, these are based on researcher observation and classification of video views of surrounding traffic. A six- level classification scheme has been used to classify exposure points and conflicts. Light (A) means free-flowing traffic, medium (B) means flowing with some restrictions (owing to the presence of other vehicles), and heavy (CâF) means var- ious degrees of restricted traffic flow. Table 4 shows these three groupings with heavy listed first as it is the highest-risk condition. In the table, notice the disproportionately high risk for heavy traffic density, equivalent risk for medium density, and lower risk for light traffic. The odds ratio of conflicts to expo- sure for heavy traffic (levels C, D, E, and F) compared with lighter levels (A and B combined) is 5.9, indicating that inci- dent risk is about six times greater in heavy traffic. Also notice, however, that the majority of conflicts still occurred in light traffic, even though relative risk was lowest. About half of all LTCCS truck-crash involvements occurred on urban roads, although only 28% were cited as having a âtraffic factor.â LTCCS trucks were at-fault in 45% of their multivehicle crashes in urban areas, versus only 33% in rural areas (Knipling and Bocanegra 2008). Trucks were also more likely to be at-fault in crashes where traffic was a factor, per- haps related to blind zones around trucks. In the project survey, both safety managers and other experts were asked to rate the driving practice âAvoid urban rush hours and other heavy traffic situations.â As with other practices, they rated the safety value of the practice on a seven- point Likert scale, from â3 to +3. The practice received a mean rating of +1.7 from safety managers, making it one of the highest-rated practices. The other-expert mean rating of +1.2 was near the middle of the 11 practices rated. Event Type Traffic Density Exposure (%) (N = 1,072) Traffic Conflict (%) (N = 914) Heavy (C,D,E,F) 36 (3) 145 (17) Medium (B) 258 (24) 216 (24) Light (A) 778 (73) 543 (59) Total 1,072 (100) 914 (100) TABLE 4 TRAFFIC DENSITY: COMPARISON OF NATURALISTIC DRIVING EXPOSURE DATA TO RISK DATA
19 Most of the case study interviewees regarded traffic density as a major factor in crash risk. Carrier G uses truck routing software which in its algorithms considers traffic characteris- tics in the vicinities of delivery locations. Carrier J, located in upstate New York, monitors New York and surrounding state traffic alerts daily to warn drivers of congestion. In a research partnership with the American Transporta- tion Research Institute (ATRI), the FHWA Office of Freight Management and Operations has developed the Freight Per- formance Measures (FPM) program. FPM (www.freight performance.org) provides extensive freight travel speed data for the U.S. highway system. Initial analyses have been of speeds and travel time reliability on five major U.S. freight corridors: Interstates I-5, I-10, I-45, I-65, and I-70. Travel speed data have been collected from more than 500,000 oper- ational trucks equipped with GPS-based automatic vehicle location equipment. Trucks are assigned an anonymous iden- tification number to maintain the confidentiality of truckers and trucking companies. The system receives position (lati- tude and longitude) and time and date data from trucks at reg- ular intervals to provide data for the travel speed analysis. Trucks that stop (e.g., for refueling, deliveries, or rest) are excluded from the calculations. An FPM service (âFPMwebâ) allows carriers and other users to obtain information on travel speeds and delays for any given place and time along 25 Interstate highways: I-5, I-10, I-15, I-20, I-24, I-25, I-26, I-35, I-40, I-45, I-55, I-65, I-70, I-75, I-76, I-77, I-80, I-81, I-84, I-85, I-87, I-90, I-91, I-94, and I-95. Users may generate Geographic Information System (GIS) maps, detailed analyses of individual corri- dors, or broader analyses across corridors. One ATRI study (Short et al. 2009) used FPM data to iden- tify the 30 worst freight bottleneck locations in the United States. This was based on FPM calculations of hourly and total âFreight Congestion Valueâ for these locations. âFreight Congestion Valueâ was defined as the freight vehicle popula- tion times the average vehicle miles per hour below free flow (i.e., 55 mph). This was calculated hourly and in total. The study did not include crash counts, but the evidence cited in this report suggests the same locations would be high-crash- risk as well. A more recent analysis lists 100 such sites in descending order (ATRI 2010). Across the 100 sites, the aver- age nonpeak-to-peak congestion ratio was 1.20. The obvious benefit of avoiding congestion delays is the time savings. But is it the greatest benefit? An Ameri- can Automobile Association study (Meyer 2008) does not squarely address the question posed, but does provide a per- spective on the overall costs of congestion versus those of crashes. The study compared the costs of crashes to the costs of congestion (for all vehicle types) by calculating a per- person cost for crashes and multiplying it by the population figures in the same U.S. urban areas studied by TTI, as described previously. Crash costs were based on FHWA comprehensive costs for traffic fatalities and injuries. Per capita congestion costs varied directly with city size. Per capita crash costs varied inversely with city size. Among all U.S. cities in the analysis, average per capita congestion cost (in 2005 dollars) was $430. Per-person crash costs in those same cities was $1,051. Thus, for the urban populations, crashes cause more than twice the economic loss (and asso- ciated harm) as does traffic congestion. EFFICIENT SCHEDULING: OPTIMAL TIMES FOR SAFE TRAVEL Consider the ebbs and flows of vehicle traffic within the 24 hours of each day and the 7 days of each week. Almost all of us adapt our driving patterns to those variations in traffic density in an effort to travel quickly and efficiently. The speed paradox described in chapter one suggests that when we seek smooth, fast travel, we also find relatively safe travel. This in turn suggests that evening and overnight driving would be safest because traffic is lightest at these times. A counter- argument is that night driving is inherently riskier because of driving in darkness, the greater likelihood of driver fatigue, and the greater presence of impaired motorists on the road- ways. In the LTCCS, 62% of truck driver asleep-at-the-wheel crashes occurred in the 2-hour period between 4:01 a.m. and 6:00 a.m. This is well known to sleep researchers as a âcirca- dian valleyâ (Knipling 2009). Alcohol use by other motorists is another major nighttime risk. One analysis found that more than one-third of fatal carâtruck collisions during the overnight hours involved an alcohol-impaired car driver (Blower and Campbell 1998). A strong majority of large-truck crashes and incidents occur during the daylight hours. Here are some percentages for day- light (including dawn and dusk) crashes and traffic conflicts: â¢ All 2008 police-reported crashes involving large trucks: 79%; â¢ 2008 fatal crashes involving large trucks: 68% (FMCSA Analysis Division 2010); â¢ LTCCS CT crash involvements: 73%; â¢ LTCCS ST crash involvements: 90%; and â¢ CT naturalistic driving incidents (Hickman et al. 2005): 75%. Unfortunately, crash and incident data alone do not provide satisfactory answers to the day-versus-night question. Crash databases precisely document crash times, but they have no corresponding exposure base to serve as a denominator for generating relative crash rates by hour-of-day. Naturalistic driving studies do provide exposure data based on onboard recordings of driving times and on randomly selected âexposure points.â In the same CT naturalistic driving data cited earlier (Hickman et al. 2005), only 59% of driving was during daylight, versus 75% of incidents. The odds ratio for incident occurrence during daylight versus darkness was
about 2.1 (0.75 / 0.59 Ã· 0.25 / 0.41 â 2.1), indicative of greater risk during daytime. A time-of-day function based on the same data found the lowest incident rates to occur during the overnight hours, whereas the highest were during the after- noon hours (Knipling et al. 2005). Naturalistic driving studies may be challenged, however, based on the concept that they capture many more non-crashes (e.g., hard-braking events) and very minor crashes (e.g., curb strikes) than serious crashes. Based on a review of naturalistic driving data, crash data, and two different mileage by time-of- day exposure sources, Knipling (2009) reached the following tentative conclusions regarding large-truck crash rate by time- of-day: â¢ Overall, the large truck fatal crash rate per VMT appears to be roughly constant across the 24-hour day. â¢ Nighttime fatal crashes are more likely to involve driver fatigue or alcohol use (by other motorists), whereas day- time fatal crashes are more likely to involve traffic inter- action errors. â¢ Nonfatal injury and property-damage-only crash risks are generally higher during the daytime hours and lower at night. â¢ The hours between 6:00 p.m. and 2:00 a.m. appear gen- erally to be the safest travel times for large trucks. â¢ Overall crash risk rises in the early morning hours after 4:00 a.m. owing to the âone-two punchâ of a circadian low period and morning rush hour traffic. In contrast to these findings, project survey respondents strongly favored day driving over night driving. The follow- ing are the mean Likert scale ratings for questions on this topic. The seven-point Likert scale for these questions ranged from â3 (Strongly Reduces Fleet Safety) to +3 (Strongly Improves Fleet Safety). â¢ Maximize day driving to avoid driver fatigue and other nighttime risks: â Safety managers: +1.5 â Other experts: +1.2. â¢ Maximize night driving to avoid daytime traffic: â Safety managers: â0.4 â Other experts: â0.7. Larger carriers are more likely to analyze their crashes in relation to exposure factors such as time-of-day. For example, case study Carriers C, D, and E all conduct such analyses. Several interviewees regarded overnight driving, particularly in the early morning hours, as more risky than day driving. More definitive research on this issue is needed because of the contradictory findings and because of the potential safety benefits of reliable guidance on this issue. This research might include fleet-based studies in which both crash incidence and exposure can be classified by hour-of-day, or studies of free- ways or major trucking lanes. 20 Although its emphasis was not on traffic safety, a large urban pilot test on truck deliveries has demonstrated huge time and cost savings from shifting day deliveries to nighttime. The Research and Innovative Technology Administrationâfunded pilot test arranged for participating carriers to make off-hour deliveries, instead of their regular day deliveries, to retailers and other receivers in New York City. The pilot test found that off-hour deliveries increased travel speeds by up to 75% and reduced unloading times at receiver sites by about 70%. It also reported a sharp reduction in parking tickets and fines, which for daytime deliveries averaged more than $1,000 per month per truck for participating carriers (NYC DOT 2010). The draft project report by Rensselaer Polytechnic Institute (Holguin- Veras et al. 2010) gave no comparative crash data, but sug- gested that crash rates are lower with off-hour deliveries. Driver feelings of safety may be less, however, because of their personal safety concerns about night deliveries in a large city. AVOIDING ADVERSE WEATHER Adverse weather is an obvious source of risk in driving and, when extreme, can be a direct cause of crashes. A U.S.DOT report (Rossetti and Johnsen 2008) argues that the role of adverse weather in truck crashes is actually increasing relative to the overall truck-crash problem. That is because weather- related fatal truck crashes have declined less slowly over the past decades than have non-weather-related fatal crashes. There are a large number of different weather and climate fac- tors that can affect CMV transport safety. Some of these are listed in the textbox. The percentage of large-truck crashes affected by adverse weather depends on the criteria used for âweather-related.â In 2008, about 15% of truck crashes occurred during rain or other ânon-clearâ weather condition. About 19% of fatal crashes and 22% of nonfatal crashes occurred on wet surfaces (FMCSA Analysis Division 2010). In the LTCCS, where causal factors were more closely scrutinized, 14% of truck crash involve- Weather Effects on CMV Transport Rainâloss of traction and control, delays Snow and iceâdelays, loss of traction and control, tire damage from chains, ice on tops of vans Thunderstorms and tornadoesâdirect damage, impaired visibility, loss of control Temperature extremesâstresses on vehicle components, perishable cargo High windsâvehicle instability and blowovers, especially vans Wet pavementâloss of traction and control, road spray Hurricanesâdirect damage, road closures Floodingâroad closures, weak braking Slides (snow, mud, rock)âcollisions, delays Source: Rossetti and Johnsen (2008).
21 ments had weather as an associated factor, but less than 1% of truck at-fault crashes were assigned a weather-related CR. Juxtaposing these percentages suggests that adverse weather contributes to many more crashes than it causes directly, and this was exactly the finding of a Canadian study that selected both primary and contributing crash causes (Gou et al. 1999). In that study, âslick roadâ was deemed the primary cause of only one of 195 truck crashes, but it was the secondary or ter- tiary cause of 23 crashes (12%). Enlightened risk management requires an estimate of the relative crash risk in adverse weather or on slick roads com- pared to dry conditions. A 1980 NTSB report estimated the relative risk of fatal crashes on wet versus dry roads (for all vehicles) to be about four. In the LTCCS, 18% of CT crash involvements occurred on wet roads, versus 11% of ST involvements. A comparison of the LTCCS CT wet roads percentage (18%) to a naturalistic-driving wet roads expo- sure estimate of 9% (Hickman et al. 2005) suggests a relative risk closer to 2. Questions 1 and 2 of both project surveys asked respon- dents to select the most important (Question 1) and least important (Question 2) general factors affecting truck crash risk. Overall, âweather and roadway surface conditionsâ was considered less important than driver characteristics (both enduring and temporary) and roadway characteristics/traffic conditions (e.g., road type). Only vehicle characteristics were rated as less important. Another question on the surveys asked about the impor- tance of avoiding adverse weather and slick roads. On the seven-point Likert scale from â3 (âReduces Fleet Safetyâ) to +3 (âImproves Fleet Safetyâ), this practice was given a mean rating of +1.8 by safety managers, making it one of four top choices. The mean rating by the other experts was +1.3, fourth among 16 driving situations and practices. FHWA is developing a Road Weather Management Pro- gram that, when completed, will provide information on current and predicted road surface and weather conditions to highway users. Information sources will include fixed road sensors and instrumented vehicles, including commercial vehicles in regu- lar operations. Its outputs will be âdecision support systemsâ to aid road maintainers (e.g., snow removal operations) and travelers. At this writing, the project is just beginning to deter- mine what information motor carriers need and how best to provide that information. It is studying the economic impacts of weather on motor carriers. More information on this FHWA research and development program is available at http://www. ops.fhwa.dot.gov/weather/index.asp. The Clarus Initiative (FHWA 2007) is a joint effort of the U.S.DOT Intelligent Transportation System (ITS) Joint Pro- gram Office and the FHWA Road Weather Management Pro- gram. The word âclarusâ is Latin for clear. The Clarus Initiative is a multiyear effort to develop and demonstrate an integrated surface transportation weather observation data management system, and to establish a partnership to create a Nationwide Surface Transportation Weather Observing and Forecasting System. State DOTs and other road operators are the programâs principal target users, though the program may provide prod- ucts and services to transport companies in the future. Weather and road condition information may be provided to travelers by means of navigation and route optimization services. This would make those services more dynamic and responsive to predicted conditions. VEHICLE SIZE AND CONFIGURATION Large trucks come in different sizes and configurations, and are selected for their uses primarily on the basis of productivity and practicality. A logical question is whether these sizes and configurations are optimal from the safety perspective. Larger trucks might be safer if their use results in fewer trucks on the road and, therefore, less exposure to risk. Smaller trucks might be safer if they are individually less likely to be in crashes and if their crashes are less severe because of their smaller size dif- ferential compared with other vehicles. Answering this safety question is extremely difficult, however, because of several major variables confounding comparisons. A pair of questions on both the safety-manager and other- expert survey forms asked respondents to state the general directions of their views on larger versus smaller trucks. Both questions asked for ratings on a seven-point Likert scale rang- ing from â3 (âReduces Fleet Safetyâ) to +3 (âImproves Fleet Safetyâ). The two questions, intentionally worded to state opposite strategies, were: â¢ Use fewer, larger trucks (e.g., multitrailer trucks) when possible. â¢ Use more, smaller trucks (e.g., single-unit trucks) when possible. Safety-manager respondents assigned each strategy low posi- tive mean ratings, suggesting perhaps that either strategy could be good for safety when used appropriately. The mean ratings were +0.6 and +0.2 for the two respective items. For other experts, the two means were +0.4 and â0.3, respectively. There were wide ranges of responses for both questions for both respondent groups. The following two subsections present statistics and evi- dence to help frame the issue of whether shifts in the truck size and configuration mix would enhance safety. They do not provide definitive answers, however, because of various complexities to be discussed. Single-Unit Versus Combination-Unit Trucks Large trucks are defined as those with gross vehicle weight ratings (GVWR) of greater than 10,000 lb; 80% to 90% of
large-truck crashes involve heavy trucks with GVWRs of greater than 26,000 lb. The two major large truck configu- rations are combination-unit trucks (typically tractor- semitrailers) and single-unit trucks (also called straight trucks). CTs are typically in long-haul service, whereas most STs are short-haul. STs are more numerous (6.8M ver- sus 2.2M U.S. registrations in 2008), but on average they are driven only about one-fifth the mileage of CTs (12,362 mi per ST versus 64,764 mi per CT; FHWA 2010). As predom- inantly long-haul vehicles, CTs are more likely to be driven on Interstate highways and in rural areas between cities. Statistics for 2008 from the FHWA VM-1 vehicle mileage table (FHWA 2010) indicate that 49% of CT mileage was on Interstates, versus 21% for STs. The percentage of miles on rural roads (of all types) was 55% for CTs versus 43% for STs. Most CTs are operated across state lines. This makes them, their drivers, and their carriers subject to the Federal Motor Carrier Safety Regulations (FMCSRs). Many STs are used intrastate and thus are subject only to state regulations. Many CTs are employed in multiday trips, whereas most STs are day- use vehicles, with their drivers returning home at the end of each shift. ST driving is more likely to involve regular physical activities other than driving; indeed, many STs are primarily work-support vehicles rather than cargo-delivery vehicles. These are among the operational differences making CT-ST comparisons problematic (Knipling and Bocanegra 2008). One might automatically assume that CTs would have higher crash risks than STs because they are much larger, articulated (making them vulnerable to jackknifes and more vulnerable to rollovers), and permit less visibility around the truck. Actually, overall CT crash rates are considerably lower than those of STs, probably the result of the differences in road type exposure cited earlier. However, when CT crashes occur, they are generally more severe. Table 5 presents a com- posite analysis of CT and ST crash rates, severities, and âbot- tom lineâ crash costs per mile based on three sources. Mileage data are from FHWA VM-1 statistics, vehicles in crashes from NHTSA (2010), and mean crash severities (expressed as cost) from Zaloshnja and Miller (2007). Because the compo- nent source statistics are from disparate sources, the derived statistics need to be considered rough estimates. Table 5 shows the CT crash rate per VMT to be consider- ably lower than that of STs. CT average crash severity (average 22 monetized value of all crash injuries and damage), however, is considerably higher. From the crash rate and average per-crash costs, one can calculate average crash costs per mile for each vehicle type. As it turns out, the opposite-direction differences in crash rate and crash severity cancel each other out almost exactly. Derived crash costs per truck-mile are almost equal for the two vehicle types based on these sources. An earlier study (Wang et al. 1999) used 5 years (1989â 1993) of NHTSA General Estimates System crash statistics, FHWA mileage statistics, and the same approach to crash severity estimates to derive comparisons very similar to those presented previously. In the Wang study, the ST crash rate was 28% higher than the CT rate, but the CT average crash severity was 24% higher than that of STs. Average crash costs per VMT were 9.7Â¢ for CTs and 10.0Â¢ for STs. For both analyses, these costs represent the harm to all parties involved in all crashes, regardless of fault. They do not represent financial losses to carriers, which are much lower. These crash costs-per-mile derivations might suggest that hauling freight by means of CTs has less overall risk, because CTs have much greater capacities. The same total freight could be hauled by fewer vehicles. However, the CT and ST risk statistics are based on quite different road risk exposures, because CTs are used much more on lower-risk roads. There- fore, no generalized conclusions may be drawn. Fleets may want to consider the factors in developing their risk avoid- ance strategies, but the decision likely comes down to their own individual operational needs. Higher-Productivity Vehicles The previous questions extend to the use of trucks larger than standard CTs. Higher-productivity vehicles (HPVs) are those with GVWRs of more than 80,000 lb, the maximum size of standard tractor-semitrailers. This report does not address the many policy and regulatory issues surrounding the use of HPVs, but will address the question from the perspective of car- riers deciding how best to haul cargo within current regulations. Longer combination vehicles (LCVs) are HPVs with more than one trailer. LCV tractors may pull two or three trailers with different configurations, subject to different restrictions. Spe- cific LCV configurations include six-axle tractor-semitrailers, Metric Truck Type VMT1 Trucks in Crashes2 Truck-Crash Involvements Per 100M VMT Cost Per Crash3 Crash Costs Per Truck Mile ST 83,951 M 190,000 226.3 $56,296 12.7Â¢ CT 143,507 M 190,000 132.4 $97,574 12.9Â¢ Sources: 1FHWA VM-1 Statistics; 2NHTSA (2010); 3Zaloshnja and Miller (2007). Other statistics derived. TABLE 5 COMPOSITE ANALYSIS OF ST AND CT CRASH RATES AND COSTS
23 Rocky Mountain doubles, triple-trailer combinations, and Turnpike doubles. The United States and Canada have com- plex sets of rules and limitations regarding use of LCVs and other HPVs. In the United States, LCVs generally are per- mitted in western states but not in most eastern states. A more detailed discussion of HPV configurations and rules for use is beyond the scope of this report. Individual carriers, how- ever, often have some flexibility in the size and configura- tions of trucks they employ. From a carrier perspective, the most compelling HPV rationale is efficiency. A 2008 comparative study by ATRI analyzed HPV versus conventional CT efficiency under var- ious weight, travel, and load scenarios (ATRI 2008). In one comparison, ATRI found that moving 1,000 tons 500 mi with conventional 80,000-lb tractor-semitrailers would require 42 trips and 3,889 gal of fuel. Using a Rocky Mountain double weighing 120,000 lb would require just 27 trips, a 36% reduction, and 3,215 gal of fuel, a 17% reduction. Envi- ronmental benefits from reduced carbon emissions parallel the fuel savings. Safety benefits would arise from requiring fewer vehicles and trips to haul the same amount of freight, thus reducing exposure to crash risk. An Australian study (Moore 2007) found the HPV crash rate per freight ton-mile to be less than one-half that of regular CTs. Two Canadian studies (Tardif and Barton 2006; Montufar et al. 2007) looked at HPVs classified as LCVs. Both con- cluded that LCVs offer both productivity and safety benefits if their operations are closely and intelligently controlled. The Alberta Infrastructure and Transportation study (Montufar et al. 2007) analyzed Alberta LCV crashes over a 7-year period and determined crash rates and LCV crash risk factors. They compared LCV crash rates per VMT with those of light vehi- cles and three other truck configurations, including standard doubles. LCVs had the lowest overall crash rate of all the vehicle types examined. This meant an even greater advan- tage over other truck configurations in crash rate per ton-mile, because the LCVs carried more cargo. The Tardif and Barton (2006) study reviewed the use of Turnpike Doubles in Canada. They found the Turnpike Double incident rate for seven large fleets to be 0.24 incidents per million km (equivalent to 0.39 per M VMT), compared to an overall CT rate of 0.46 incidents per million km (0.74 per M VMT). The report cited other data indicating that doubles have similar or better crash rates than tractor-semitrailers in similar operations. Both Canadian groups suggested superior driver training and qualifications as reasons behind the observed lower crash rates for HPVs com- pared to conventional CTs. HPV drivers tend to be more senior and have superior safety records. They generally receive higher pay than drivers of CTs in comparable operations. HPV crashes do have higher severity potentials than those of other trucks because of their weight and number of trail- ers. Using LTCCS statistics, Zaloshnja and Miller (2007) found the average crash harm (including both human and material components) to be considerably higher for multi- trailer truck crashes than for those involving one-trailer CTs. Taken together, this and the earlier studies do not permit a âbottom lineâ determination of the overall crash risk of HPVs relative to conventional CTs. Earlier in this section, the results of two survey questions on the relative safety of larger versus smaller trucks were covered. The safety-manager survey also included a question on whether carriers actually used higher-capacity vehicles, and how users rated them for safety effectiveness. Only 17 of 77 respondents reported using HPVs in their fleets. User rat- ings of their safety effectiveness were generally high, how- ever; the mean rating was 3.9 on a five-point scale. Using Full Vehicle Load Capacity Regardless of a truckâs or busâs legal load capacity, it would appear to make sense to use the vehicleâs full capacity rather than to operate partially empty vehicles. The benefits of oper- ating vehicles at full capacity are much like the benefits of reducing empty miles, as discussed earlier. Fully loaded trucks are less likely to experience wheel lockups and associated jackknifes and other loss-of-control incidents (Moonesinghe et al. 2003; Knipling 2009). Loaded vans are less likely to be affected by crosswinds than empty ones (Rossetti and Johnsen 2008). On the other hand, heavier vehicle loads increase stop- ping distances (Clarke et al. 1991), thus potentially increasing rear-end and forward-collision risks. Heavier loads also raise vehiclesâ centers of gravity slightly, adding to rollover risks (Moonesinghe et al. 2003). The âbottom lineâ probably favors full loads, though the benefits cannot be stated categorically. ONBOARD COMPUTERS AND MOBILE COMMUNICATIONS Commercial vehicle onboard computers and mobile communi- cations offer a wide range of potential applications for opera- tions and safety. Many of these applications are beyond the scope of this report. Most notably, this report does not address collision-avoidance systems, such as forward-collision warn- ings, lane-departure warning systems, and side object detec- tion systems. It also does not address the technical details of onboard computers and wireless communications systems. The term telematics comprises onboard sensors, networks, soft- ware, GPS, and wireless communications, which are becoming commonplace in todayâs commercial vehicles (Strah 2009). Much of the MCES focuses on mobile communications used to support various operational efficiencies (Belella et al. 2009). The focus here will be on those specific telematic applications mentioned by motor carriers in project surveys and interviews, which relate to both operational efficiency and safety. These were discussed primarily with regard to safety benefits, though some concerns were expressed about safety losses owing to driver distraction. Commercial vehicles have been equipped with âcomput- ersâ for about 2 decades, at first in the form of electronic
engine control modules. Qualcomm and other companies introduced satellite communications, which created the poten- tial for remotely accessing onboard computer data. Cellular- based communication systems now have similar capabilities to satellite systems (Strah 2009). Current systems are becoming complex and comprehensive fleet monitoring and manage- ment tools. Telematic functionalities previously offered by third-party vendors are increasingly being offered by truck manufacturers at the time of purchase. Systems allow central, real-time viewing of a vehicleâs map location, moving speed, engine speed, battery and fuel status, and trip history. Vehicle component (e.g., brake, tire) condition monitoring is also available (to be discussed in the section Monitoring Vehicle Condition). The system can be programmed to flag any trou- ble indicator, whether it relates to vehicle functioning or driver behavior. A report by Aarts (2008) for the Organisation for Economic Co-operation and Development (OECD) and the International Transport Forum (ITF) describes various rapid, ongoing, efficiency-related changes in commercial vehicle technolo- gies, logistics, and infrastructure. The report anticipates expected effects of mobile communications systems and other vehicle telematics on road transport efficiency, fuel use, pollutants, infrastructure, and safety: The use of mobile communications tracking systems can enhance the security and efficiency of commercial vehicle operations by providing information about asset locations and a direct means of communication between carrier personnel and drivers. By closely tracking vehicles and assets, opportunities for cargo and vehicle theft can be reduced. Additional benefits include poten- tial improvements in delivery service and asset utilization through vehicle location and routing information. Human resource man- agement and worker productivity can be enhanced by carriers receiving more accurate status and arrival time information on shipments. Increased visibility into this information can expedite deliveries and help to ensure on-time performance to customers. The OECD/ITF report indicated that initial cost to carriers was the principal barrier to greater market penetration of telematics, which it placed at about 35% of the European mar- ket. The FMCSA Technology Division (2010a) has published a technology product guide to wireless communications and related technologies. The guide explains much of the technol- ogy and its applications in carrier management. The product guide lists 12 system vendors. Any discussion of in-vehicle technologies needs to consider the problem of driver distraction from such devices. Ergonomic issues relating to the safe use of telematic systems by drivers are beyond the scope of this report. Suffice it to note that driver dis- traction from cell phones, other in-vehicle devices, and from other sources has been recognized as a major cause of crashes. The U.S.DOT has conducted two national summits on dis- tracted driving. Almost 1,600 U.S. companies have adopted distracted-driving policies, and a new federal law prohibits tex- ting by commercial drivers. An FMCSA naturalistic driving study of distraction risk found that drivers interacting with a dispatching device while driving had an odds ratio of 9.9 for 24 incident involvement and that their eyes were off the road an average of 4.1 s/6 s of driving (FMCSA 2009). On project surveys, respondents were asked to rate the safety benefits of using onboard computers and of using mobile communications. As with similar questions, respon- dents were presented with a seven-point Likert scale ranging from â3 (âReduces Fleet Safetyâ) to +3 (âImproves Fleet Safetyâ). The overall mean for safety-manager respondents with regard to onboard computers was +1.1. The mean rating for onboard communications was just +0.6, with 17 negative ratings among 78 respondents. With regard to actual use of onboard computers, just over half of respondents used them (41 of 74). Among those who did, their mean safety effectiveness rating was 3.9 on a five- point Likert scale. For mobile communications, 58 of 78 respondents used them; users assigned a mean safety effec- tiveness rating of 3.6. More detailed questioning would be needed to sort out respondent views on specific benefits from these systems and on potential concerns about their use. In the case studies, the comments of the Case Study C inter- viewee were pertinent to the question of driver distraction from onboard devices. The official interviewed believed that the key challenge was to communicate and provide infor- mation without causing distraction. The company does not use driver-accessible general-purpose computers in its cabs because of the potential for distraction. It does use commu- nications and navigational aids, but without providing visual displays when vehicles are in motion. An electronic device converts any text sent to drivers to voice when the vehicle is moving so that driversâ eyes are not diverted from the road. TEAM DRIVING This section and the next three address topics that were not included in the original project work plan or in the safety man- ager surveys but have been added to the discussion because they were mentioned by carrier safety managers in project surveys (chapter three) or in interviews for the case studies (chapter four). The four topics are team driving, EOBRs, fuel economy and safety, and vehicle condition monitoring. Although each topic merits more detailed coverage, brief discussions are provided here to round out the discussion of safety-relevant carrier efficiencies. Team driving is an efficiency practice because a team- driven long-haul truck legally can be moving almost continu- ously during an extended trip. One driver can rest during the other driverâs driving period so that no stoppages are required by HOS rules per se. Of course, stops are still required for fuel, food, personal hygiene, and breaks away from the vehicle. FMCSA has estimated that 9% of truck VMT is driven by team drivers. This would mean that, at any given time, about 17% of drivers are involved in team operations (FMCSA and
25 ICF International 2007). Team driving has several important safety advantages, but also some disadvantages. Knipling (2006, 2009) summarized the advantages and disadvantages as follows: â¢ Safety advantages: â Presence of second person in vehicle reduces unsafe driving acts; â Social interaction sustains alertness; â Reduced continuous-time driving (time-on-task); â Sleep and breaks can be on demand; â Greater regularity of sleep; â Greater overall sleep time; and â Familiar and secure sleep location. â¢ Safety disadvantages: â Poorer quality of sleep in moving vehicle; â Shorter durations of principal sleep period; and â Possible greater fatigue at start of trip (if driver sleep schedules are not coordinated). Team driving appears to have lower crash risk than solo driving. A naturalistic driving comparison of team and solo driving (Dingus et al. 2001; FMCSA 2002) found the team driver incident rate to be less than one-half of the solo driver rate. Team drivers had far fewer driving misbehaviors such as speeding and tailgating. Although sleep quality was lower for team drivers (because they were sleeping in moving vehicles), sleep times were longer. The team driver rate of high-drowsiness incidents was just one-fourth the solo driver rate. Team drivers were much less likely to push themselves to the limit and therefore avoided high-drowsiness incidents. Team driving presents management and operational chal- lenges, however. Many carriers would utilize team driving more extensively if they could better meet these challenges. Questions and challenges relating to team driving involve trip planningandrouting,vehicle features (i.e., sleeper berths), team driverrecruitingandassignments,dailyworkand driving sched- ules (e.g., use of split sleep), and safety management practices. Married couples generally make the best and happiest driver teams because of the close driving and living conditions. No item addressing team driving was on the safety-manager survey form, but one was added to the other-expert form. Respondents rated the item âMaximize use of driver teams for long haulsâ on a â3-to-+3 Likert scale. The mean rating assigned by 31 respondents was +0.8, equal to the grand mean rating for the 16 items rated. Only one of the 11 case study interviewees explicitly mentioned team driving as a safety strategy; the Carrier B interviewee noted his com- panyâs support for it as both a safety and efficiency measure. ELECTRONIC ONBOARD RECORDERS The topic of EOBRs was not within the original scope and was not included in project surveys. However, EOBRs are dis- cussed briefly because they were cited by several case study interviewees as aids to both efficiency and safety. EOBRs monitor commercial driver HOS compliance by maintaining a readable electronic time record of vehicle movement (driving) and of time duration since the dayâs initial driving. EOBRs are used voluntarily by a growing number of CMV fleets, but are currently required only for those carriers with the worst histo- ries of HOS noncompliance. FMCSA is considering extending the EOBR requirement to a larger percentage of noncompliant carriers. This discussion addresses only the efficiency and safety management benefits of EOBRs, not regulatory issues surrounding them. Apart from EOBRsâ effects on HOS compliance and driver fatigue per se, they are seen by some fleets as an aid to more efficient operations and to safety management. Eight of the 11 case study carriers (see chapter four) either were using EOBRs or were transitioning to them, and most considered them beneficial for both safety and efficiency. Because EOBRs automate driver log-keeping, they save driversâ time, streamline records and compliance management, and provide a means for safety oversight of drivers through quick identifi- cation of noncompliant drivers. EOBRs also facilitate load assignments in larger fleets by identifying drivers with suffi- cient time available for the loads. One corporate vice presi- dent for safety interviewed in conjunction with the study noted that EOBRs help carriers to âdraw a line in the sandâ in their interactions with customers. Customers might assume that paper logs allow HOS compliance flexibility, whereas EOBRs reinforce a need for absolute compliance. Shackelford and Murray (2006) found EOBR benefits included improved fuel consumption monitoring and fuel tax compliance, quicker tabulation of driver mileage and loads, easier tracking of vehicle and engine wear, real-time vehicle location monitoring, and better communications and dispatch- ing. The study even reported improved driver morale. FUEL ECONOMY AND SAFETY Another carrier efficiency factor with safety implications is fuel economy. Several interviewees believed their efforts to improve fleet fuel economy had safety benefits. Maximizing fuel economy has cost-reduction benefits for companies and the environmental benefit of reducing emissions. Devices and driving practices improving fuel economy also reduce vehicle wear, tire wear, and maintenance costs (Smith and Roberts 2007). Improved fuel economy is achieved in large part by changes in vehicle speed and driving style. These changes in turn produce safety benefits such as reduced driver stress, crash likelihood, and crash severity. Two primary approaches to improving fuel economy that have concomi- tant safety benefits are speed-limiting vehicles and monitoring individual driver fuel consumption. Speed Limiters Speed limiters, also called speed governors, are devices that limit the top powered speed of vehicles. Modern truck enginesâ
electronic control modules are easily programmed to limit top powered speeds to some set point. And, because âexcessive speedâ was the most frequent proximal cause of truck crashes in the LTCCS, some might regard speed limiters as a top- priority crash countermeasure. One must realize, however, that speed limiters cannot prevent most truck crashes arising from excessive speed. That is because most instances of âexcessive speedâ occur on lower-speed roads and at speeds below top freeway speeds (e.g., 65 mph). Moreover, speed limiters would not slow the downhill speeds of trucks. Speed limiters would, however, reduce both the likelihood and severity of crashes involving trucks and buses traveling at speeds greater than the top freeway speeds. CTBSSP Synthesis 16 (Bishop et al. 2008) examined the safety impact of large-truck speed limiters. The project included a safety-manager survey based on a convenience sample, similar to the current study survey. In the MC-16 proj- ect survey, 56% of respondents indicated speed limiters were either âsuccessfulâ or âvery successfulâ in reducing crashes. Speed limiter users believed that limiters were either âsuc- cessfulâ or âvery successfulâ in reducing tire wear (44%) and increasing fuel economy (76%). Almost 96% of respondents believed that speed limiters had no negative effects on either their companyâs safety or productivity. Speed limiters are already required on trucks in European Union countries and in Ontario and Quebec in Canada. In the United States, NHTSA and FMCSA have proposed federal regulations for speed-limiting heavy trucks, and the matter is under rule-making consideration. Much of the trucking industry favors mandatory speed limiters on large trucks (ATA 2006), and many companies are adopting them volun- tarily (Bishop et al. 2008). Reduced crashes are the primary rationale, but other reasons include lower fuel and mainte- nance costs, reduced emissions, and longer tire life. The proj- ect survey included no items on speed limiters, but some respondents and case study interviewees commented on their efficiency and safety value. No quantitative crash rate reduc- tions were reported, though one earlier study accessed in the literature review found that trucking firms with firm speed limit policies had crash rates 30% lower than those of their peers (Dammen 2005). Several of the case study interviewees stated that their trucks were electronically speed limited, usually at 65 mph. Those mentioning speed limiting also stated that they moni- tored driver fuel use, as discussed next. Monitoring Driver Fuel Economy A more direct method for improving fuel economy is to mon- itor fuel use for individual drivers and trips. A capability for onboard fuel consumption monitoring is commonplace in todayâs trucks. Advanced, electronically controlled engines automatically monitor fuel consumption. Many EOBRs also monitor fuel consumption (Shackelford and Murray 2006). 26 Fuel economy can also be monitored conventionally with- out special onboard capabilities. Internationally, an initiative called ecodriving (www.ecodrive.org) is promoting greater fuel economy for all vehicles. Ecodriving focuses on driving style. Its âFive Golden Rulesâ are: 1. Shift up as soon as possible. 2. Maintain a steady speed. 3. Anticipate traffic flow. 4. Decelerate smoothly. 5. Check tire pressure frequently. With the exception of Rule 1, all of the rules for improved fuel economy are also rules for safer driving. Ecodrivers are âsmooth operators.â They learn to adopt a smoother driving style, âglidingâ through traffic, shifting to the highest gear possible, and avoiding rapid accelerations and decelerations. Drivers learn to look down the traffic stream as far ahead as possible to predict and react smoothly to changes and inter- ruptions in traffic flow. This defensive, anticipatory driving style also serves to reduce crash risk. In the United Kingdom, more than 13,000 heavy vehicle operators have received ecodriving training, with a reported average fuel savings of 10% (SAFED 2010). Symmons and Rose (2009) described an ecodriving training program in a trucking fleet that reduced fuel consumption by 27%, gear changes by 29%, and brake applications by 41%. Another ecodriving training and monitoring program reportedly resulted in a 13% fuel savings at Setz Transport Company (IRU 2003). The Setz program involved fuel consumption monitoring, positive recognition for drivers showing improve- ments, and remedial training for those not showing improve- ment. A 2007 TRB paper (Zarkadoula et al. 2007) described a successful pilot test of ecodriving involving urban bus drivers in Greece. The SAFED (2010) bus and coach web page reported 12% fuel savings, a 40% reduction in gear changes, and a 60% reduction in âsafety-related faults,â although the latter was not defined or explained. The term âecodrivingâ is commonly used in North Amer- ica, but many fleets monitor fuel use for individual drivers. Fuel use may be the basis for driver rewards, positive recog- nition, or discipline. Many companies use the same vehicle monitoring capabilities to measure hard-braking events, which are themselves correlated with fuel consumption and crash risk. Almost all of the project case study companies (see chapter four) monitor individual driver fuel use and compo- nent behaviors such as hard braking and speeding. For exam- ple, Carrier J (Small Charter Bus Service), a small charter bus company, has equipped all of its motor coaches with a multi- function electronic monitoring system. The system provides onboard safety monitoring (OBSM) of driving behaviors and electronic HOS logs. The OBSM system records and reports top speeding time (i.e., above a specified top speed), highest observed speeds, hard-braking events and rate, fuel use, and other driving efficiency and safety indicators. The system
27 generates a âDriver Report Cardâ for each trip. The com- panyâs safety director reported that driver acceptance of the monitoring was good and that the drivers even âmake it a competitionâ to see who can earn the best scores. Zuckerman (2009) described various fleet efforts to train drivers to decrease their fuel consumption, with associated safety benefits. One fleet owner interviewed identified âhigh acceleration and jack-rabbit starts-and-stopsâ as the principal targets for remediation. Minimizing speeds per se is less impor- tant than minimizing rapid accelerations and decelerations. Training and other practices suggested included: â¢ Use of speed limiters to eliminate the highest speeds; â¢ Instrument panelâmounted fuel-use displays to give drivers feedback on fuel use; â¢ Training drivers to resist the urge to speed up for yellow lights, but rather to anticipate light changes and coast slowly to stops; â¢ Use of cruise control; â¢ Monthly analysis of individual driver fuel use and driving patterns; â¢ Rather than discipline, emphasis on rewards and recog- nition for best performers; and â¢ For large fleets, extending the training and rewards up the line to fleet managers and supervisors. No item addressing fuel economy monitoring was on the safety-manager survey form, but one was added to the other- expert form. Respondents rated the item, âMonitor fuel econ- omy for individual drivers and provide feedbackâ on a â3-to-+3 Likert scale. The mean rating assigned by 31 respondents was +0.7, near the grand mean rating for the 16 items rated. MONITORING VEHICLE CONDITION Automatic monitoring of vehicle condition was not included in the project survey but was cited by several case study inter- viewees as a growing application with both safety and effi- ciency benefits. Mechanical maintenance deficiencies are far more common in large trucks than in light vehicles because of their larger number of components and their more continuous use. In the LTCCS, 40% of crash-involved trucks had some vehicle-related deficiency or malfunction, and the presence of such deficiencies was strongly associated with fault in crashes. Mechanical failures were much less frequently a prin- cipal cause, however. Overall, about 4% of LTCCS truck crash involvements were assigned a CR of vehicle mechani- cal failure, with another 2% as a result of cargo shifts. As discussed in the section on preventive maintenance ear- lier in this chapter, almost all successful motor carriers prac- tice systematic PM. By regulation, drivers are required to make pre-trip vehicle inspections each day. To supplement these measures, numerous automatic vehicle condition monitoring technologies are penetrating the fleet. These can provide con- tinuous monitoring and feedback to drivers and recordings to onboard electronic data recorders. Wireless transmission of vehicle condition data to roadside enforcement is an emerging capability, with potential efficiency benefits to commercial transport and safety benefits to everyone. Such monitoring can potentially include brake adjustment and condition (the most common vehicle-based problem in inspections and crashes), tires, lighting, and vehicle weight. Tire pressure monitoring exemplifies truck vehicle condi- tion monitoring, and is relevant to both safety and efficiency. In the LTCCS, 1.1% of at-fault truck crashes were caused primarily by tire failure. The percentage was much higher for STs (2.2%) than for CTs (0.7%; Knipling and Bocanegra 2008). Poor tire condition is the second most common vehi- cle source (behind brakes) of violations in truck roadside inspections. Many of these crashes and violations could be avoided by proper tire care and regular inspection. The most common cause of tire failure is underinflated tires, which can become overheated and have excessive sidewall flexing (Freund et al. 2006; Knipling 2009). A 2003 study of large- truck tire inflation (Kreeb et al. 2003) found that many fleet operators do not perform the regular tire pressure maintenance recommended by tire manufacturers. The study reported that: â¢ Approximately 7% of CMV tires tested were under- inflated by 20 psi or more. â¢ Only 44% of tires were within 5 psi of their specified target pressure. â¢ Tire-related costs were the single largest maintenance expense for CMV fleets, averaging about 2Â¢ per mile or about $2,500 for an annual 125,000-mile operation. â¢ Improper inflation raised total tire-related costs by $600 to $800 annually per tractor-trailer combination. â¢ Improper tire inflation increased annual procurement costs for new and retreaded tires by 10% to 13%. â¢ Larger fleets are generally more systematic and rigor- ous than smaller fleets with regard to tire pressure and other tire maintenance. An FMCSA safety technology product guide, available on its website, describes various types of tire pressure monitoring systems (TPMS) available from nearly 20 vendors (FMCSA Technology Division 2010b). These devices also save pre-trip inspection time, improving operational efficiency. Flanigan (2010) reported that approximately 5% of fleets use TPMS. The small system penetration was said to be the result of fleet concerns about system reliability, maintenance costs, and ini- tial costs. This situation may be changing rapidly, however. In Transport Topics, Reiskin (2010) reported a survey of man- agers of large U.S. fleets finding that 43% use TPMS. Wide- spread use of TPMS by large fleets may portend greater penetration across all fleets. Challenges associated with the use of TPMS include proper training for maintenance staff, consistent and correct use of data from the systems, and disciplined inspections and track- ing of the sensor systems themselves to ensure that they do not
add to overall vehicle maintenance workload (Van Order et al. 2009). A recently published fleet test of onboard brake perfor- mance and tire pressure sensors (Van Order et al. 2009; Flani- gan 2010) used three different TPMS on 36 tractors and 39 trailers in two fleets. Preliminary results from one fleet and 4.6 million miles of travel found the use of TPMS to be asso- ciated with slower tire wear and 1.8% better fuel economy. Complete and final project results are pending at this writing. GENERAL RELATIONSHIP BETWEEN EFFICIENCY AND SAFETY The previous sections have presented findings relating to spe- cific tactics to increase both carrier efficiency and safety. What about the general relationship between efficiency and safety? In the aggregate, do the various practices cited previously add to both greater efficiency and greater safety? Do the same car- rier practices and processes that foster efficient operations also foster safe operations? The project did not measure either the efficiency or safety of any fleet, so it cannot provide defin- itive evidence. A question on both the safety-manager and other-expert survey forms, however, asked respondents about the general relationship. High majorities of both categories of respondents selected the choice, âHighly efficient carriers tend also to be more safe than other carriers.â Only 2 of 77 safety managers believed that such carriers were less safe. None of the 31 other-expert respondents selected this choice. In an article entitled âHigh-Performance Work Systems and Occupational Safety,â Zacharatos et al. (2005) surveyed 138 manufacturers (including chemical, automotive, and con- struction) that were members of the Industrial Accident Pre- vention Association of Ontario. The surveys contained 124 Likert-scale questions on high-performance management prac- tices and the extent to which employees practiced and were 28 committed to high performance. Two companies participated in more in-depth surveys of individual frontline industrial supervisors. âGenericâ high-performance management prac- tices were associated with both company financial perfor- mance and with safety measures. Figure 5 shows some of the correlational relationships among various corporate measures. Of most interest here is the relationship between a corporate high-performance work system and corporate safety climate. Striving for high organizational performance may have neg- ative impacts on safety if it results in excessive demands for productivity. Efforts to increase productivity in factories, for example, can result in higher accident rates if workers are per- forming tasks too fast for safety (Blum and Naylor 1968). Caird and Kline (2004) looked at job demands and driver safety behaviors among 150 work drivers in a large western Canadian company. The study found that organizational demands encouraging high work speed were associated with higher lev- els of driver fatigue, more errors, and more moving violations. Larger carriers have advantages over smaller companies and owner-operators in creating operational efficiencies with potential safety implications. Many proactive and systematic management practices are practiced more widely among larger carriers than among smaller ones (Stock 2001; Corsi and Barnard 2003). Larger companies may also feel less pres- sure to push productivity and delivery schedules to unsafe levels. In Australia, Mayhew and Quinlan (2006) interviewed 300 long-haul commercial drivers to assess economic pres- sure, driver workload, and occupational safety and health. Owner-operators reported worse health and safety than did drivers in small fleets and, especially, those in large fleets. Structured interviews revealed a connection between eco- nomic pressure (e.g., month-to-month dependence on income from loads) and negative health and safety outcomes. Owner- FIGURE 5 Interrelationships among five organizational characteristics among 138 industrial manufacturing organizations. Source: Zacharatos et al. (2005).
29 operators were less likely to seek medical treatment for injuries or illnesses, often citing financial pressures as the reason. Drivers working for large firms felt more secure about report- ing sickness and injuries and in taking time off for them. Can one therefore demonstrate a link between overall motor carrier performance and safety? Corsi et al. (2002) compared the financial performance of 656 carriers with their safety per- formance to determine whether there was a link. The study controlled for carrier features such as size, revenues, average load, and average haul. Financial data were obtained from the ATA corporate annual report database, and carrier safety rat- ings were retrieved from SafeStat. The 656 companies in the study were mostly large, prominent carriers. Carriers with satisfactory ratings (553 of the 656) had an average 3% profit margin. The 103 companies not rated satisfactory (96 condi- tional + 7 unsatisfactory) had average operating losses of 4%. On a second financial measure, return on investment (ROI), satisfactory fleets had a 5% average ROI. This compared to a negative 2% ROI for non-satisfactory (conditional or unsatisfactory) fleets. Though there were exceptions, safer companies generally had higher profits. The study contained no measure of efficiency per se, but the findings of this large study with regard to safety and profitability suggest a simi- lar link between safety and efficiency. The following are safety-manager survey comments relat- ing to ways in which efficiency can relate to safety, positively or negatively: â¢ You must develop a safety culture from the top down. You must be willing to make investments in technology to promote safety. â¢ Executive and management involved in all levels of safety and compliance. â¢ Accountability for safety. â¢ No one practice but a culture. â¢ Efficiency and safety must be used in conjunction, and not considered âstand-aloneâ initiatives. The following are similar comments from other experts: â¢ Managers of well-managed operations pay attention to all aspects of their operations. â¢ Some efficiency measures in dispatch may result in non- rested drivers being given long runs that will result in fatigue. â¢ Some standard practices with respect to vehicle type, operations, and schedules may promote efficiencies but not fit a driverâs ergonomic needs, circadian rhythm, and temperament, which often lead to increased risk. â¢ Need to adopt a systems-based approach, applying a model or framework such as the Haddon Matrix. â¢ To manage efficiency implies an organizational struc- ture that can also manage risk. Two of the case study interviewees commented on the general issue of operational efficiency in relation to safety. Carrier Aâs management structure is designed to strengthen the link between efficiency and safety. Division operational managers are also safety managers for their divisions and are evaluated based on both operational and safety success. Carrier Jâs safety manager had had previous experience in pickup and delivery (P&D) operations. He felt that intense monitoring of driver delivery times by management sometimes forced drivers to work too fast and cut corners on safety.