Video Surveillance Uses by Rail Transit Agencies (2011)

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.

34 CHAPTER FOUR ADMINISTRATIVE CONSIDERATIONS IN THE USE OF ELECTRONIC VIDEO SURVEILLANCE INTRODUCTION As surveillance technology advances and its use becomes commonplace throughout the transit industry, agencies are being presented with an expanding list of places that it can be employed. The decisions are not made in a vacuum, and once a decision to rely on video surveillance is made, a num- ber of questions arise. An important question is whether the cameras will be monitored in real time or used solely for forensic investigation. Once this is decided, additional ques- tions must be answered. For instance, if the system will be viewed in real time, will it be always monitored or only dur- ing operating hours, and who will do the monitoring (i.e., police/security, rail operations, or some combination of these staffs). Whether viewed in real time or later, questions need to be answered about how and where images will be stored and who will have access to them. Additional ques- tions may arise surrounding whether those on the transit system’s property (patrons, employees, or even trespassers) should be informed that their actions are being monitored by video surveillance. This chapter relies on questionnaire responses and the literature review to provide a snapshot of how agen- cies have answered some of these questions. It also pro- vides examples of how such decisions impact policies and procedures. DECIDING WHERE TO INSTALL VIDEO SURVEILLANCE The use of surveillance in the United States is not as widespread as it is in the United Kingdom, but it has been steadily expanding. It is not unusual for newspaper readers around the nation to see stories about their cities increas- ing their reliance on cameras for a number of crime pre- vention efforts. New York City, Chicago, Baltimore, and Pittsburgh are only a few of those whose mayors have spoken frequently on the issue, and many smaller cities have turned to cameras without the fanfare and publicity of these larger municipalities (Figure 4). Announcements of transit agencies’ expansion of their surveillance networks also receive local attention from the media. FIGURE 4 The New York City Police Department posts signs on local streets indicating the presence of security cameras. This sign was across the street from a Manhattan subway station. Photo courtesy of Dorothy M. Schulz. To provide some guidance on why certain surveillance installations and placements were made, agencies that had installed surveillance cameras on fewer than 76% of their stations, station platforms, or shelters were asked why cer- tain locations were covered and others were not. Using a five-point scale ranging from least to most important, the two most important factors in determining which stations, station platforms, or shelters had surveillance or on which it might be installed were “high disorder/crime rate” and “funds available to retrofit” (Table 7). Decisions on where and when to employ electronic sur- veillance may be influenced by patron expectations, which have changed considerably with the current focus on safety and security. For instance, when, on December 7, 1993, Colin Ferguson shot 23 people—6 fatally—on an LIRR train, no one asked why the railcar lacked surveillance video. The response to this event may have been tempered by Fer- guson’s immediate capture by an off-duty transit officer, but more recent crimes on transit properties that are not captured

35 ies, including the local or regional transit systems. Two of the most vocal officials have been Chicago’s Mayor Richard M. Daley and New York City’s Mayor Michael Bloomberg, whose cities contain, respectively, the second-largest and largest U.S. transit systems. Differences in the current status of surveillance-related issues by CTA and NYCT highlight not only political issues, but also problems that may occur with vendors, particularly in retrofitting aging heavy rail transit systems. In May 2010, Mayor Daley and CTA President Richard L. Rodriguez announced that by May 31 at least one or more surveillance cameras would be installed in all 144 CTA sta- tions and that nearly 3,000 cameras would be installed by the end of the year (“Mayor Daley …” 2010). The announcement came less than a year after Daley appointed Rodriguez CTA president and encouraged him to focus on improving the safety on the system. Crime on CTA had increased slightly in 2008 and early 2009, and when he was appointed Rodri- guez noted that cameras were installed on every bus and that he hoped to have them at all train stations with 18 months. The CTA has received $22.6 million in DHS funds since 2006, a portion of which is being used to expand its network of cameras. Although the transit system has also invested approximately $19 million of its own funds in the project, the importance of DHS funding was highlighted by Chi- cago Transit Board Chairman Terry Peterson, who noted that the DHS grants have allowed CTA to make “significant upgrades to the security and surveillance network” (“Mayor Daley...” 2010). CTA began adding cameras in 2002; in Jan- uary 2011 Amy Kovalan, CTA’s chief safety and security officer, announced that CTA would install cameras on about half its rail cars based on an existing DHS grant and had a “verbal agreement” from DHS to pay for installation on the remaining cars. An installation schedule was not provided (“CTA to add security cameras to trains” 2011). The Chicago Police Department, whose transit division provides policing for the CTA, estimated that cameras had played a role in more than 4,500 arrests since 2006. Rodriguez also noted the importance of the cameras to the CTA Con- trol Center, which views real-time video to assist in passen- ger safety by monitoring and managing service disruptions and by providing the City’s Office of Emergency Manage- ment and Communications the ability to communicate with police, fire, emergency response, and CTA personnel during incidents. In the past, although he has not provided specific numbers, Mayor Daley has stated that Chicago’s network of public and private surveillance cameras is “the largest in the United States” (Spielman 2010). The announcement of the expansion of the surveillance network included information that all new railcars would arrive with cameras pre-installed and that a pilot program would be undertaken to determine the feasibility of retrofitting existing vehicles. on video leave the agency subject to criticism. For instance, in May 2010, NYCT was criticized after an encounter in a Greenwich Village subway station that was not captured on video led officials to admit that almost half the cameras in the subway did not work. Ironically, the absence of video played no role in the case. The 19-year-old suspect arrested in the deaths of two other men was released when a grand jury refused to indict after deciding he had acted in self defense (Eligon 2010, p. A13). TABLE 7 MOST/LEAST IMPORTANT REASONS FOR INSTALLING VIDEO SURVEILLANCE ON STATIONS, STATION PLATFORMS, AND SHELTERS Reason Most Important Least Important High disorder/crime rate 4 1 Local demands/politics 3 3 Enhance perceived cus- tomer safety 2 1 Funds available to retro- fit existing stations, station platforms, or shelters 4 1 New stations, platforms, or shelters designed to accommodate surveil- lance devices 2 0 Other (none specified) 0 0 Similarly, agencies that reported that fewer than 76% of their railcars were equipped with surveillance technology were asked to indicate the most and least important reasons for equipping some vehicles and not others. Table 8 indi- cates the number of agencies for whom choices were most and least important. TABLE 8 MOST/LEAST IMPORTANT REASONS FOR EQUIPPING RAILCARS WITH VIDEO SURVEILLANCE TECHNOLOGY Reason Most Important Least Important High disorder/crime rate 1 2 Local demands/politics 2 4 Enhance perceived cus- tomer safety 4 0 Funds available to retro- fit existing railcars 2 2 New railcars equipped at purchase 2 3 Other (none specified) 0 0 Although only a small number of transit systems indi- cated that politics played a role in installation of surveillance cameras, a number of major cities’ mayors have been vocal in their support of video surveillance throughout their cit-

36 receive considerable financial support for equipment and for employee training and terrorist awareness programs from the federal government. As indicated by the responding transit agencies, cur- rently the major funding source for surveillance systems is the DHS Transit Security Grant Program, followed by fund- ing from the FTA. APTA has also increased its presence in transit security, and both it and the FTA have published a number of studies of best practices that were the basis for many of the directives issued by the TSA in 2004 to public transit agencies (Jenkins and Butterworth 2007). DHS/TSA has awarded grants for planning, training, equipment, and other security enhancements, in addition to providing other services to transit agencies. Some grant programs have been used to undertake risk assessments and bolster emergency response capabilities. The largest percentage of the avail- able funds, though, is used for employee training and for the purchase of surveillance equipment. Prior to the creation of DHS in the wake of the 9/11 terror- ist attacks, FTA was the most common source of funds for purchasing equipment; Maier and Malone (2001) reported that 14 of their responding agencies, which included both bus and rail systems, received funds from the FTA grant program, 9 relied on state grants, 6 on local funds, 6 on inter- nal funding sources, and 1 on an unspecified source. They noted, though, that about one-third of the agencies used a combination of sources to fund their purchases (2001, pp. 23–24). The FTA helps transit agencies fund security proj- ects by providing financial assistance and by requiring that agencies spend 1% of their urbanized area grant program funds on security improvements. These funds are available to jurisdictions with populations of 50,000 or more for use for capital investments, operating expenses, and transporta- tion-related planning. The existence of DHS funding has had a direct influence on rail transit expenditures for security. For this study, agen- cies were asked to indicate the percentage of their funding that came from a number of sources, including the FTA grant program, DHS, state grants, municipal grants, agency funding, funding or grants from surveillance equipment vendors, or any other sources. Agencies were not asked to provide dollar amounts; they were asked only to indicate the percentages of funds from each source. Table 9 indicates the number of agencies that listed receiving more than 50% of their funds for surveillance expenditures from any one of the choices provided and those that indicated that 100% of their funds came from any one source. Agencies that did not receive at least 50% of their funds from a sole source but from a variety of the sources are not included. One light rail system than anticipates rev- enue service beginning in 2011 received an equal percentage of funding from FTA and DHS (50% received from each). New York’s mayor—and its transit system—have been less successful in plans to increase video surveillance in NYCT’s subway system. Although its size and the age of the system combine to make NYCT unrepresentative of tran- sit agencies, its experiences are instructive of problems that can occur, albeit on a smaller scale, for all transit systems. In April 2009, the MTA, which oversees the NYCT, was sued by its video surveillance contractor, Lockheed Martin. The company alleged that the agency’s interference relating to its $300 million contract to install a network of digital cameras had prevented it from completing work begun in late 2005. Two months later the MTA countersued, claiming Lockheed Martin had “bungled” the antiterrorism program that was intended to link 2,000 subway cameras into an intelligent video surveillance command center. At the time, the MTA stated that only about 1,400 of the 1,750 cameras were installed and that few were working. The basis of its countersuit was that the system had failed repeatedly during tests and that Lockheed Martin had falsely claimed the work was progressing even though about $250 million had been spent (Namako 2009). Regardless of the claims and counterclaims, some of the cameras’ inability to capture video was attributed to the tran- sit environment, where heat, water, and electrical problems slowed the job’s progress. These are all factors that may limit attempts to retrofit a century-old transit system to accept mod- ern technology (Rivera and Grynbaum 2010). The lawsuit is pending; because of this, MTA and its constituent agencies declined to reply to this synthesis’s questionnaire. Many factors may influence the decision to install a sur- veillance network in all or part of the transit system. In the case of Tri-Met, the route played a role because concern centered on its 5.5-mile Airport MAX line, which travels from downtown Portland to Portland International Airport. Since it began revenue service in September 2001, video sur- veillance has been employed along the line, but because the Airport MAX terminus was in close proximity to the air ter- minal, the FAA requested that no train be unattended at the airport. In addition to adding security patrols to the airport station during all operating hours and checking all trains for unattended items, Tri-Met installed surveillance cameras at the airport station. This also illustrates the expanded role of the federal government in local decisions since 9/11. Using grant funds, Tri-Met also installed cameras on all 78 MAX trains, at stations with elevators, and at a number of parking garages (Eder 2005, p. 1927). FUNDING VIDEO SURVEILLANCE SYSTEMS Because transit agencies are local entities, each needs to purchase surveillance equipment independently of other transit agencies. In the aftermath of terrorist attacks on transit systems worldwide, U.S. transit systems began to

37 Valley Metro in Phoenix, a case study agency, was the only agency whose surveillance network was funded solely with agency money. As a new transit system, it was not eligible for DHS grants but will be able to compete for such funds now that it is fully operational. TABLE 9 SOURCE OF 50%/ALL VIDEO SURVEILLANCE FUNDING Source of Funds for Video Surveillance System 50% or More All Funds FTA Grant Program 11 2 DHS 10 4 State Grants 1 1 Municipal Grants 0 0 Agency Funding 1 0 Vendor Funding/Grant 0 0 Other 0 0 Because agencies were promised anonymity, analysis of the table is general. The two agencies that received 100% of their funds from FTA are recently opened systems. The agencies that received 100% of their funds from DHS include Amtrak (which is federally funded overall), two commuter, and one light rail agency. The agencies that received 50% or more of their funding from either FTA or DHS do not fall into easy categorization with the possible exception that a far larger number of FTA-funded agencies are newer systems in areas of the country that are less likely to be seen as major terrorist targets, although there were exceptions to either of these descriptions. With few exceptions, the agencies that received at least 50% of their funds from DHS were estab- lished transit agencies in urbanized areas. The three agencies that received more than 50% of their funding from sources other than FTA or DHS are equally difficult to categorize. The agency that received all its fund- ing from state grants is a new commuter rail system and the one that received 50% of its funding from state grants is an established light rail system located in a different state. Finally, the agency that funded more than 50% of its sur- veillance-related costs from its own funds is a large, long- established eastern seaboard agency. Results received in answer to this question underline that the sources of funding for surveillance are limited even if the dollar amounts are considerable. Only a few agencies reported receiving funds or grants from vendors, and these generally cover 10% or less of their costs. Although com- muter rail agencies are regulated by FRA and would be eli- gible for FRA financial support for surveillance initiatives including under the FHWA’s Highway-Rail Crossing Pro- gram, none indicated this as a source of funds. The responses overwhelming reinforce the importance of external funding for purchasing and upgrading surveil- lance systems. Although transit systems must in effect com- pete against one another for the DHS funds, the amounts of money available are larger than from any other single source. For instance, in May 2010 the DHS announced that it would release almost $790 million in Preparedness Grants for nine federal programs. The Transit Security Grant Pro- gram was to receive $253.4 million, plus an additional $150 million provided through the first and second American Recovery and Reinvestment Act funding provisions. Also included in the total of $790 million was $20 million to Amtrak; $14.5 million to the Freight Rail Security Grant Program for critical freight infrastructure projects centered on transportation of hazardous materials; and $11.5 for the Intercity Bus Security Grant Program, which is available to fixed-route intercity and charter bus companies for security planning, facility upgrades, and vehicle and driver protec- tion. Other funds are allocated to other areas of transporta- tion infrastructure, including ports and terminals (Kronfeld 2010). Despite these large amounts of available money, politicians and the DHS’s own Office of the Inspector Gen- eral have consistently urged DHS to expand its effective- ness in the area of mass transit and passenger rail. A recent report, though, focused primarily on nonmonetary aid (Chunovic 2010). Reinforcing the close ties between terrorism prevention and detection and risk management, DHS disburses funds only to agencies that have relied on its mandated method- ology to complete a risk assessment. This requirement has resulted in a number of agencies that previously had not completed risk assessments undertaking them to be eligible to apply for funds. Basic eligibility to compete for funds is based on the Urban Areas Security Initiative list and the National Transit Database; eligible applicants are listed as part of the annual guidance published to assist agencies in completing the requests. DHS has further divided agencies into two tiers. Tier I is composed of transit agencies in the eight highest-risk urban areas as determined by DHS; Tier II consists of all other eligible transit agencies. Agencies are effectively competing against one another. Applications are evaluated by panels composed of federal employees who score the projects based on a number of criteria, including the agency’s risk group score, the project’s effectiveness group score, the project’s potential for risk mitigation (which includes cost-effectiveness, feasibility, timeliness, and sus- tainability), regional collaboration if required, and the agen- cy’s offering of a cost share. Projects are ranked and funded in order until the funds are exhausted. All information and forms are available on the DHS website, as is information on Tier I-eligible agencies, the allocations for each Tier I area, and designation of Tier II areas and eligible agencies. Some agencies are in more competitive areas than oth- ers. MBTA and MARTA, for instance, are the only Tier I

38 agencies in their states. Although ACE, a case study agency, is part of the San Francisco Bay Area Tier I, it competes for funds in that group ($19,873,038 in FY 2010) with much larger, higher profile agencies in the Bay Area’s Regional Transit Security Working Group. All other case study agen- cies are designated as Tier II. Funding issues explored by this synthesis centered on purchase. The synthesis did not pursue costs and issues pertaining to the operation or maintenance of surveillance cameras, including the related costs of hardware or software that are required to maintain the surveillance system in an operational state. These costs are considerable and insuf- ficient maintenance of an existing surveillance system can contribute to negative publicity about an agency and may influence how claims of loss or injury are adjudicated either by internal claims officers or by courts. Technical studies of the actual operations of surveillance systems by rail agen- cies might assist them in determining whether their initial purchases are being supported internally by policies and procedures that maintain the equipment properly. These studies might also consider how internal decisions impact the effectiveness of the video surveillance system and the transit system overall. PERCEIVED EFFECTIVENESS OF VIDEO SURVEILLANCE Decisions about where to install cameras are influenced by an agency’s goals, available funding, and, sometimes, con- cerns of the political entity to which the agency is linked. Intertwined in each or all of these decisions is the perceived effectiveness of the surveillance network. Although effec- tiveness might be difficult to define in the context of these overlapping but possibly contradictory goals, agencies were asked to indicate how effective their surveillance systems were in achieving a number of goals. The two major reasons for employing video surveillance monitoring of locations were for crime/vandalism prevention and accident investi- gation. Respondents could select as many or as few of the choices that pertained to their agency (Table 10) TABLE 10 REASONS FOR EMPLOYING VIDEO SURVEILLANCE Reason No. Crime/Vandalism Prevention 39 Fare Collection View/Dispute Mediation 20 Other Complaint Resolution 32 Accident Investigation 39 Employee Monitoring 18 Other 10 Again using a five-point scale, agencies were given a list of the most common purposes of a video surveillance sys- tem and asked to rate from most effective to least effective whether they believed this goal was met on their system. Table 11 indicates the number of times an agency listed a reason as most effective and the number of times it was rated as least effective. TABLE 11 MOST/LEAST EFFECTIVE USE OF VIDEO SURVEILLANCE Reason Most Effective Least Effective Crime Prevention/ Vandalism 7 2 Fare Collection/Dispute Mediation 4 4 Other Complaint Resolution 7 0 Accident Investigation 11 0 Employee Monitoring 0 0 Other 1 1 The perceived effectiveness of a video surveillance sys- tem can depend on a number of administrative issues beyond purchase, installation, and maintenance of the equipment itself. Transit systems need to address many operational issues when considering upgrading an existing surveillance system or installing a new one. Survey respondents provided information on how a number of these are addressed, includ- ing policies on monitoring, recording, and archiving images, and whether patrons and/or employees are notified of the presence of video surveillance technology. MONITORING VIDEO CAMERAS—WHEN AND BY WHOM As the number of cameras increases, questions have arisen as to whether they will be monitored in real time (some- one watching the cameras as things are happening) or will be viewed after the fact (looking at images after an incident occurred). A related decision is who will monitor the cameras and for what hours whoever is assigned will view them. The times that video cameras are monitored differed considerably. Twenty-two agencies indicated their cameras were monitored at all times (24 hours a day, 7 days a week), and six reported that cameras were never monitored. Eight responded that cameras were viewed only during hours of transit operations, while 11 indicated they used a configura- tion that was not easily summarized but met their agencies’ needs (Table 12).

39 transit policing was established by 1979, it may simply stem from the initial view of surveillance as primarily a patron traffic control and rail operations tool rather than a law enforcement tool. It may also simply reflect a continuation of past practices. VIDEO SURVEILLANCE SYSTEM FEATURES Many of the questions pertaining to the technological fea- tures of an agency’s surveillance system did not receive replies or received replies that were internally inconsis- tent. As with all questionnaires, it is difficult to determine why some questions are answered and others are not. Many respondents were vague about when surveillance was intro- duced on their system. This, in combination with the few responses to a request to provide the name of the surveil- lance vendor, supports a tentative conclusion that managers responsible for daily operation of the surveillance system are less interested in the technical specifications of their systems than in its day-to-day use and reliability. With this caveat, it can only be stated that most agencies rely on digital rather than analog systems, and that more than a third described their surveillance systems as combined or in transition from analog to digital. Particularly given the current attention paid to video analytics, most of the existing surveillance systems were described by respondents as having what today would be considered relatively low-tech features (Table 14). TABLE 14 VIDEO SURVEILLANCE SYSTEM SPECIAL FEATURES Special Feature No. 24-hour Recording 36 Auto Emergency Digital Transmission 4 Secondary Power Source 18 Auto-start 8 Low Light Resolution 16 VIDEO ANALYTICS As video surveillance has proliferated two new issues have emerged: perception overload and the expanded use of sen- sors in conjunction with cameras or as stand-alone tools to protect vital areas. Both rely on advanced technology that a number of agencies are introducing into their video sur- veillance networks. Even in agencies that assign personnel to monitor images in real time, the rise in interest in video analytics is based on the realization that most surveillance systems produce far more images than it is possible for view- ers to absorb. The use of video analytics (“smart” or “intel- TABLE 12 WHEN VIDEO CAMERAS ARE MONITORED When Monitored No. 24 Hours a Day, 7 Days a Week 22 During Hours of Service Only 7 Another Configuration 7 Not Viewed 9 Note: Two agencies checked multiple responses. Although some agencies checked multiple responses, making it difficult to fully interpret the responses, the larg- est number of agencies indicated that their cameras are under constant monitoring and the smallest number indicated that the cameras are not viewed at all in real time. The six agen- cies that reported that cameras were never monitored in real time stated that the decision was based on the cost involved in having personnel assigned to this function, although two also indicated that other unspecified factors played a role in their decisions. Once an agency decides that cameras will be monitored, it must decide who will monitor them (Table 13). Although the synthesis did not delve into many areas of personnel decision- making, such as whether labor agreements were a factor, the agencies that reported that cameras were not viewed listed cost as the major reason for this decision. In most agencies, rail operations personnel are assigned to monitor system operation during all hours train are running and sometimes even when they are not. Because cameras serve a number of non-law enforcement purposes, having rail operations per- sonnel monitor video systems is consistent with the camera’s overall roles in safe operations of the rail network. On other systems, where there is a full-service 24/7 police agency that monitors emergency telephones and responds to incidents on the transit system, having those individuals monitor the sur- veillance network is also consistent with their roles. TABLE 13 PERSONNEL MONITORING VIDEO CAMERAS Who Monitors No. Police/Security Personnel 10 Rail Operations Personnel 8 Combined 22 Note: Not all agencies responded. Without more information, for instance, whether union agreements played a role in the decision, whether the initial purpose of the video system played a role in the decision, or whether existing communications networks were used to activate the surveillance system, it is difficult to general- ize as to how a combined network came to be the preferred method. Recalling that the earliest federal report noted that

40 ligent” video) attempts to provide a solution. At its most simple, smart video can be defined as video that thinks for you. It not only collects data, but is capable of analyzing the data; for instance, in addition to merely filming individuals on a crowded platform, smart video would identify and focus on persons who act suspiciously and would alert those in the viewing room to turn their attention to the monitor display- ing this particular action. Scientists who are studying it and vendors who are mar- keting it refer to smart video as the next-generation of elec- tronic video systems. Such systems rely on algorithms to profile behavior based on how people usually behave in cer- tain environments and then picking out those whose behav- ior is different from others or inappropriate for the location or situation. These systems take into consideration changes in lighting conditions, an important factor for rail facilities and parking locations, and can track people as they move from one camera to the next. Understanding and Using Analytics Video analytic systems analyze data to improve tracking. They are programmed based on what people can be antici- pated to do. If normal behavior can be anticipated, abnormal behavior can be made to stand out. For instance, to track an individual at an airport, the system is provided with infor- mation on the routes people are likely to take. The system understands and absorbs that most people go from the air- port entrance directly to the ticket check-in area, most likely then to check the flight information board, and from there to security checkpoints. Because it is designed to detect behav- ior that differs from the norm, the analytic-based system is intended to pick up someone who follows no logical pat- tern through the facility. This could be someone who stops and then starts moving again in an erratic pattern. It could be someone who seems to linger in front of doors that are alarmed or marked “employees only,” suggesting that the person might be considering whether it is possible to enter without detection. Although some transit agencies are making use of video analytics, introducing smart video into the rail environment presents a number of challenges. The major challenge is anticipating patterned behavior. This is more difficult in a transit facility than at an airport, where the most people are boarding or alighting from a plane. At a large urban transit facility, people may be shopping, walking through the station to avoid city streets in bad weather, dining at one of the facil- ity’s sit-down or fast-food restaurants, or doing any number of things that do not involve taking a train. This is less likely to occur at a small, suburban light rail station, where virtu- ally all those on the platform are apt to be waiting for a train, but patterns may still be different if the Monday-to-Friday crowd is primarily commuters carrying only briefcases and the evening or weekend crowd is made up of families taking rail to a sporting event. Furthermore, even before such sys- tems have become common, the Federal Bureau of Investiga- tion (FBI) released a warning that terrorists may be one step ahead of smart video. A jihad-advocating website reportedly suggested that adherents leave suspicious bags around New York and Washington, D.C., to desensitize first responders by forcing them to respond to suspicious but harmless items left in public areas (Weiss and Mangan 2010). NJT has used federal DHS funds to install a system that is programmed to alert those who monitor the video when a suspicious activity has occurred. In a station, this might be a bag left unattended or in a particular location. Along the ROW, it might be a boat docked under a bridge (Hecker 2006). Also in conjunction with DHS, MTA-MD has been developing a smart video system in the Baltimore subway system, light rail stations, and in Maryland commuter trains (Nakanishi 2009, p. 23). St. Louis’ MetroLink combines tunnel intrusion with analytics to monitor its tracks and tunnel. The intrusion sensors indicate activity in the area while the analytics are able to determine whether the intru- sion is authorized or not (Resnick 2009). Boston’s MBTA is using smart video elements in its recently updated camera network, particularly in and around tunnels. Smart video is also a large component of the National Capital Region Rail Pilot Program involving Amtrak, as discussed in Chapter three. The use of analytics is also a feature of the surveil- lance systems of two case study agencies, Metro Transit and Valley Metro (see chapter five). Distinguishing the Usual from the Unusual Because of the greater focus on airport security than rail security, a demonstration of smart video in late 2009 at airports in the United Kingdom used footage obtained at Heathrow International Airport, where a group of scien- tists said their prototype identified potential threats that human operators would have missed (Fleming 2009). In a study conducted among Florida transit agencies, Dmitry B. Goldgof and colleagues (2009) found that few agencies were knowledgeable about analytics. The study also referred to a number of drawbacks, including an analytic system’s vulner- ability to environmental variables such as detrimental light- ing conditions and weather, both of which may lead to false alarms that could become a source of frustration for the user. Another drawback, particularly in environments where not all activities can be anticipated, was that to properly program an analytic system, events need to be predefined; events that have not been defined will not be detected (p. vi). Announcements on breakthroughs in the area of ana- lytic software appear regularly in the security and tech- nology trade press, which makes it difficult for operations managers to keep up with the changing technology. For instance, in the first 2 weeks of June 2010, researchers announced that a computer vision system that was not yet

41 Among others listed were risk assessment staff, safety and claims managers, facilities managers, legal counsel, and a variety of rail operations personnel, often in the last instance limited to supervisory personnel. A number of agencies chose not to answer this question. Although it is difficult to interpret this lack of response, it may merely indicate that the person who completed the form is not involved in this area of administrative decision-making. If this presumption is accurate, it indicates a need for policy coordination among all those with responsibility for use and maintenance of the video surveillance system. An area related to who may access images is what pro- cedures exist to ensure that only those designated with the authority actually have access. To address this area, the sur- vey asked a question on procedures that were used to main- tain a record of access (often referred to by law enforcement personnel as the chain of custody). Of the 32 agencies that indicated they permitted only designated individuals to access images, 10 had specific sign in/sign out procedures. Five agencies said that designated individuals were required to access the records only with another person present, and five indicated another control mechanism such as writing in a log. Though “only designated individuals” is likely to be sufficient for internal review, it can be anticipated that particularly in a criminal court case, a more formal sign in/ sign out policy will be required to meet chain of custody requirements. Last, there is the question of public access. Of the 41 agencies that answered the question on public access policies, 17 indicated they had none. The issues of the length of time images are retained, who may access them, and developing a formal mechanism to track access appears to be a fruitful topic for discussion at professional association meetings. Agencies with more for- mal policies that have had experience relying on their images might share information with less experienced agencies. VIDEO SURVEILLANCE AS A FORENSIC TOOL The importance of policy development surrounding use of surveillance images is directly related to its use as a forensic tool in both criminal and tort (civil) prosecutions. If agencies intend to offer images as evidence in court and in formal internal disciplinary matters, they will be asked to describe how the images are safeguarded, how they are labeled as to location and time, and what chain of custody policies ensure that the images are not tampered with and are actually the ones on which charges were based. Use of images for criminal or civil prosecutions is com- mon. Thirty-seven agencies indicated that either their own police or local police used images from their surveillance cameras in court cases. This is a large increase over Maier and Malone’s 2001 finding that 10 of 19 agencies had used ready for commercial use could provide a live text descrip- tion of video images to alleviate some of the time and labor of searching though video or image collections. Another set of researchers announced the development of software that would also save time and tedium by summarizing a whole day’s video in a few-minutes-long synopsis. In the same period, DHS announced a pilot project in conjunction with the Massachusetts Port Authority to test a system at one terminal of Boston’s Logan Airport that puts together a number of cameras to provide a 360-degree wide view and can analyze images with sufficient detail to scan for abnormal activity and for suspicious items left behind or removed (Beauge 2010; “Hebrew University Invention …” 2010; Simonite 2010). As with the use of surveillance cameras along the ROW and of sensors in conjunction with existing or upgraded video surveillance networks, analytics is a relatively new technology that will undoubtedly receive more attention from transit agencies as it becomes more readily available and as funds become available for additional research and purchase. ARCHIVING, RETAINING, AND ACCESSING SURVEILLANCE IMAGES Agency policies on archiving, retaining, and accessing sur- veillance system images differ considerably. A majority of agencies (35 of 42 respondents) archive the images from their video surveillance systems, but differences exist in the length of time images are retained. In some cases, this has to do with state laws; Florida, for instance, mandates that images from surveillance systems be retained for at least 1 month. Archiving images is only one of a series of decisions that need to be made about surveillance systems. Overall, reten- tion ranged from none at all unless something exceptional was observed or reported to more than a year, including up to 3 years in one agency. Access to images is another impor- tant policy area with organizational and legal ramifications. Although most agencies indicated that only “designated individuals” could access images, the definitions of these individuals were was not consistent. But despite the differences in policies, certain gener- alizations are appropriate. In all agencies with their own police departments, police may access the images, although in some agencies access for forensic investigation may be limited to detectives/investigators or supervisory staff mem- bers. Agencies with security departments, rather than fully empowered police departments, are more likely to limit access to supervisory personnel. When the security staff is supplied by an outside contractor, only high-level supervi- sors or agency employees who manage the security staff are likely to be authorized to access images.

42 in these areas would provide guidance to transit agencies and could preclude costly and time-consuming litigation. PATRON AND EMPLOYEE AWARNESS AND PERCEPTIONS OF VIDEO SURVEILLANCE The vast majority of agencies (31 of 41) notify patrons that surveillance cameras are in use. As with record/image reten- tion, whether to notify patrons of the presence of surveillance cameras may be strictly an administrative decision or may be based on state regulations. The U.S. Supreme Court ruled in 1967 in Katz vs. United States (389 U.S. 347) that there is no reasonable presumption of privacy in a public place. Following the reasoning of United States vs. Knotts (368 U.S. 276) in 1983 that persons traveling on public thoroughfares had no reason- able expectation of privacy, the same applies to transit facilities. In addition to meeting legal or regulatory obligations to provide signage or other notification, such as periodic announcements on the use of video, signage indicating the presence of video surveillance has been seen by many agen- cies as a way to enhance patron’s perception of safety and security. In providing examples of the signage used by a number of transit agencies, Maier and Malone (2001) noted that most include phrases such as “for your protection,” “for your safety,” for your safety and security,” or “for your safety and comfort.” Others simply stated that the vehicle was equipped with cameras or that cameras may be onboard (p. 26). Maier and Malone noted that many agencies used the words “may be recorded” rather than “is recorded” because the latter implies that cameras are always operating, which may not be accurate. This may raise legal issues if some- thing were to occur at a time when the cameras were not in operation. Similarly, if signage implies that cameras are monitored, patrons may mistakenly believe that if they have a problem, it is being viewed in real time and that someone will be dispatched to help them. Patron perception surveys could assist agencies in learn- ing more about whether the existence of surveillance sys- tems leads to less fear among riders. Surveys could also help to determine whether existing signage is properly under- stood by riders and others making use of transit facilities, yet relatively few agencies report having measured patron perceptions of security since surveillance was installed. A few who had not measured it said it had existed since the beginning of revenue operations and believed their patrons would not be able to make any comparison with how they felt without surveillance. Of the 32 agencies that answered, 12 had measured patron perceptions through surveys or other instruments; 11 of these stated that patrons reported feeling a higher sense of security (Table 15). recordings from their surveillance systems as evidence in court. The increase can be attributed to higher quality images being available from upgraded camera networks and also to courts having become more accustomed to accepting video images as evidence. The media tend to be intrigued by video evidence. Cases in which it plays a role are frequently publicized widely in local newspapers and on television stations, where the video image is often shown frequently on news programs. Two typical examples include a 2006 arrest made in con- junction with a stabbing that occurred on a GCRTA trolley after the assailant was identified based on video images. At the time, GCRTA said that videos were not viewed in real time at its command center but that drivers were trained to activate the onboard system when an incident occurred to ensure that the data was recorded over (Gural 2006). In a similar incident, in 2009, video cameras in place on an MBTA bus led to the arrest of five people who were charged with assault with a dangerous weapon (Irons 2009). In Phil- adelphia, police were able to arrest a suspect who is alleged to have attacked a SEPTA passenger with a hammer. Although other passengers ignored the assault, the suspect was later identified after surveillance video that aired on local television resulted in his identification (“Philadelphia Police Make Arrest…” 2008). In addition to indicating that video had assisting in crimi- nal prosecutions, almost as many agencies (32) reported that they had used images from their surveillance systems as evidence in employee disciplinary actions. The two ques- tions may or may not be related, but of the 39 agencies that reported whether they had seen a reduction in fraud/injury claims based on their surveillance systems, 25 answered yes and 14 answered no. Fraudulent claims may come from a number of sources other than employees, such as “ghost rid- ers” and individuals who claim to have lost items or been injured in some way on the agency’s property. The relationship between surveillance evidence in disci- plinary actions that result in a reduction of internal fraud/ injury claims appears to be fruitful area of further study. A better understanding of whether there is a relationship between the presence of surveillance cameras and employee fraud and/or discipline would be of particular value because the issue of cameras in operator/cab areas has become a prominent and controversial one and is likely to become a labor/management negotiating issue. As surveillance systems proliferate in public areas, many civil liberties groups have filed or have indicated they are planning to file lawsuits surrounding this. A review of exist- ing laws, pending litigation, and any existing model policies

43 TABLE 15 HAS AGENCY CONDUCTED PATRON PERCEPTIONS SURVEYS Perception Yes No Measured patron perceptions since surveillance was installed 12 29 Patrons indicate a higher sense of security 11 N/A Note: Figures total 41 responses; two agencies did not respond. N/A = not available. Only nine agencies had measured employee perceptions of security since the installation of video surveillance tech- nology. Similar to patrons, some employees have worked where such equipment has always existed, which would make it difficult to determine how effective a measure it is of their feelings of security at their work sites or, on the other hand, whether they believe it is there solely to monitor their productivity or adherence to work rules. All of the agencies that had measured employee perceptions said that employ- ees reported a higher sense of security. Agencies were also asked whether unions or employee representatives had been consulted in the decision to install surveillance technology; of the 31 replies, 22 said no and 19 said yes. On a transit sys- tem where virtually all operating employees are unionized, MBTA has included unions and employee organizations in its decision to install surveillance technology, which is used onboard vehicles only in passenger areas. An interesting area of study would be whether older systems or those whose employees are covered by union agreements are more likely to consult with employee representatives. Also, newer sys- tems that included surveillance as part of their initial plan- ning may not have a need to consult with employees because utilization of surveillance does not represent any change in working conditions. Quantitative data can be useful for agencies to compare their own practices with those of other agencies. Quali- tative information, which is generally provided in a nar- rative format that makes it easy to highlight details and point out lessons learned, can also help agencies to learn from others. Chapter five provides case studies to assist transit professionals who have been and will continue to make decisions on the purchase and use of surveillance systems by comparing their own situations with those of other transit systems.