Transportation Research Implementation: Application of Research Outcomes (2015)

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.

40 SESSION 4 using Research Results in Effective Ways Luis Fernando López Ruiz, Administrator of Railway Infrastructures (ADIF), Madrid, Spain Allen Biehler, Carnegie Mellon University, Pittsburgh, Pennsylvania, USA Christopher Martin, Robert Bosch Corporate Research, Pittsburgh, Pennsylvania, USA Natalia de Estevan-Ubeda, Transport for London, London, United Kingdom Jesús Rodríguez introduced the session by comment-ing that the company he works for, which has head-quarters in Europe, has its main activities in Canada and the United States. Such cross-Atlantic corporate work points to the value of cross-Atlantic cooperation in research as well. He noted that research can cover dif- ferent applications, but researchers can identify common inputs and outputs and can carry out the research in ways that make it more useful to buyers. The private sector wants to improve competitiveness, for example. Rodrí- guez explained that this session would cover four areas: railroads, roads, the interaction between modes and infrastructure, and, finally, the urban environment. For example, the Infravation program can serve as a guide for research proposals to involve industry, as well as the owners and operators of highway and roads, so as to demonstrate the potential impact of the project and the application of the demonstration results. implemenTaTion of r&d resulTs in railWay infrasTrucTure Luis Fernando López Ruiz Luis Fernando López Ruiz began by giving some background on his organization, Administrator of Railway Infrastructures (ADIF). ADIF is a public company under Spain’s Ministry of Development and administers the Spanish railway infrastructure. Railway infrastructure administration involves overseeing railroad tracks, rail stations, and goods terminals. In particular, ADIF oversees rail traffic management, capacity allocation of rail to operators, and the royalties received for the use of rail infrastructure, stations, and goods terminals. ADIF has almost 14,000 employees and manages 2,322 kilometers of high-speed rail and 13,000 kilometers of conventional rail. Fully 784 million travelers use the Spanish railway annually on 1.8 million trains that operate with 95% to 98% punctuality. ADIF’s research facilities and labs are located in numerous areas within Spain. ADIF’s Research, Development, and Innovation Policy ADIF’s research, development, and innovation (R&D&I) policy is to bring developments into production. López commented that before he became director of R&D&I, he was responsible for operations; he noted the difference in mind-set between operational people who have to solve daily problems in the short term and researchers who work on long-term projects and do not have to face immediate problems. He pointed out that this difference in mind-set helps to explain why some projects fail and others succeed. In 2006, ADIF evaluated whether, as a public company that manages tracks, it needed to have a specialized R&D department. Ultimately, the answer was yes, an R&D department was necessary, and the president issued a policy declaration outlining ADIF’s R&D&I policy. Specifically, the policy called for R&D&I to

41u s i n g r e s e a r c h r e s u l t s i n e f f e c t i v e w a y s • Control and reduce technology risks, • Give the company the ability to position itself at the forefront of technology, • Develop and maintain a technology watch, • Identify and prioritize the most appropriate mechanisms for protecting and exploiting research results, and • Carry out the transfer of technological developments. ADIF’s R&D&I Process López showed a chart of the R&D&I process at ADIF (Figure 1). The process starts with an idea and then moves to a prototype to evaluate technical feasibility. The next steps involve protection of intellectual property (IP), implementation of the technology within ADIF, incorporation of the technology into ADIF’s production process through a marketing agreement, and documentation of the experience of its use and improvement, where applicable. The final step of the process is a transfer of the technology to third parties. Next, López showed a chart of R&D&I projects undertaken. The number of projects undertaken increased steadily from three in 2005 to a high of 47 in 2011 and 29 in 2013. López pointed out that there was a gap in the number of projects started and the number completed. For example, in 2010, 44 projects were started but only 11 were completed. In 2012, 29 projects were started and 21 were completed. López noted that projects that have clear objectives end within the allotted time, but when the objectives are not clear, the projects take longer. ADIF’s R&D&I projects are classified into one of five technical specialties and one of four strategic objectives, López said. The five technical specialties are • Infrastructure, • Energy, • Control command and signaling, • Telecommunications, and • Rolling stock. The four strategic objectives are • Increasing the operational performance of the infrastructure; • Improving energy efficiency; • Increasing the rel iabi l i ty, avai labi l i ty, maintainability, or safety metrics that make up the railway transportation system; and • Developing the railway of the future. López next described the five-step life cycle of proj- ects. The life cycle begins with planning the basic require- ments of the research and deciding whether to undertake the project collaboratively or only internally. The second step is executing the project (from detailed requirements to design, prototype, testing, and validation). Final- izing the project is the third step and is accomplished by documenting the results achieved. The fourth step is undertaking an internal transfer within ADIF to pass the research results into the production cycle, and the last step is the transfer to third parties (Figure 2). FIguRE 1 ADIF’s R&D&I process. general direction of operations and construction. (Source: ADIF.) Implementation in ADIFR&D&I project Incorporation into its production process. Experience of use and improvement (where applicable) Addition to its production process after the experience of consumer use Protection Marketing agreement Transfer to third parties Prototype for technical feasibility R&D&I infrastructure Transfer Oficina Española de Patentes y Marcas

42 t r a n s p o r t r e s e a r c h i m p l e m e n t a t i o n Results of ADIF’s Research Investments Next, López discussed the results of ADIF’s R&D&I investments. The overall budget of the R&D&I public– private partnership (PPP) projects as of May 2013 was €84 million, of which €57 million came from private partners, €20.9 million came from the science system, and €5.9 million came from ADIF. Thus, for every €1 that ADIF invests, the private sector invests €9.7 and the science system invests €3.5. ADIF’s research investments have yielded numerous intangible assets, including 25 national patents, 14 international extensions (Patent Cooperation Treaty, European Patent Office), seven licensed patents, 12 current licenses, seven national utility models, two licensed utility models, two current licenses, four national industrial designs, 34 national trademarks, six EU trademarks, three international trademarks, and six intellectual property licenses. Commercialization Results López concluded with a discussion of the successful commercialization results of ADIF’s R&D&I, including the C-350 Contact Wire for railroad electrification, which was licensed to COBRA, SEMI, and ELECNOR and implemented on the international section of rail from Figueras, Spain, to Perpignan, France, and on the high- speed line between Mecca and Medina, Saudi Arabia. What the successful cases had in common was that R&D&I identified a problem and then solved it. ADIF’s production people have a say in what they want to improve so that R&D&I solves the right problems, López said. R&D&I makes prototypes and ensures they work and that the price and performance are better than before. All of the successful projects were collaborative projects. Another successful commercialization project was the optical fiber falling objects sensor. The sensor detects whether something has fallen on the track, thereby providing high-reliability detection and early warning of problems. The system has a low maintenance cost. Another successful commercialization project was the lateral wind sensor, which detects high winds and reduces the speed of trains in response. The DaVinci traffic control system, another successful project, provides centralized telecontrol over multiple systems. The reversible substation project, a fifth successful commercialization, increases energy efficiency by utilizing regenerative train braking. An electronic converter of direct current–alternating current recovers the electrical energy from braking trains and returns it to the supply network. Finally, the Ferrolinera 3.0 project was a sustainability project that developed and validated a system for charging electric vehicles by using the energy generated from the electric braking of trains. This project gives consumers an opportunity to recharge their electric vehicles at railway stations, which are abundant. In conclusion, López said that the secret to implementation is close ties between people working in operations, research, universities, and supply. implemenTaTion of r&d resulTs in operaTing roads and moTorWays Allen Biehler In contrast to López’s focus on rail, Allen Biehler’s presentation focused on roadways. Biehler described successful examples of implementation from research projects of four state departments of transportation (DOTs)—Arizona, Kentucky, Michigan, and Oregon— and identified common factors among them. FIguRE 2 Life cycle of projects at ADIF. (Source: ADIF.) PLANNING EXECUTION FINALIZATION INNER TRANSFER (ADIF) Basic requirements: Collaborative project or own investment. Detailed requirements: Design Prototype Testing Validation Results achieved . The project is deemed transferred when all or some of its results are passed to the production cycle . OUTSIDE TRANSFER Transfer to third parties .

43u s i n g r e s e a r c h r e s u l t s i n e f f e c t i v e w a y s Arizona DOT Research The project from the Arizona DOT examined wildlife– vehicle collisions that take place on Arizona roads. Vehicle collisions with elk or bighorn sheep are a serious problem in rural Arizona. Solving the problem is of interest not only to the Arizona DOT but also to the Arizona Game and Fish Department, which is why the Arizona DOT reached out to this sister agency to join the research team. Cooperation between the two agencies made use of their respective skills. The Arizona DOT could provide funding, project management, and implementation, whereas the Game and Fish Department could conduct the animal capture, GPS tracking, monitoring, and data gathering. The results of the research provided an interesting insight: elk will use underpasses, but bighorn sheep will only use overpasses, and the surface of the overpass matters to them as well. The Arizona DOT built overpasses and underpasses, which the wildlife is using regularly now. The successful factors of the project included having engaged and involved sponsorship. Although the goals of the individual stakeholders differed (safe roads versus preserving animal species) the ultimate aim of both parties was the same: to eliminate wildlife–car collisions. Another success factor was positioning the project for implementation from the start. Kentucky Transportation Cabinet Research Biehler’s second example came from the Kentucky Transportation Cabinet (KYTC), which for 25 years has had an exclusive relationship with the University of Kentucky. In particular, Biehler described the Cumberland Gap Tunnel floor settlement project. The tunnel is used by a major freeway that crosses Kentucky, but parts of the tunnel floor were sinking. This was a serious operating problem because detours around the tunnel were extremely long. The project used ground-penetrating radar to find voids and found 7,000 square feet of voids, an indication that this was a long-term problem. The voids were the result of the limestone subbase dissolving. The solution proposed filling the voids with inert granite material. The lessons learned from this project were to keep the university involved in framing all the research and to have continuous collaboration between the university and the KYTC research team. Finally, implementation was woven into the capital program. Michigan DOT Research Biehler’s third example, from the Michigan DOT, focused on longitudinal cracks in concrete box beams (i.e., cracking and joint problems). The Michigan DOT worked with researchers at Lawrence Technological University. The research delivered a series of transverse tensioning rods that used an alternative to steel—carbon fiber—as the tensioning material. The solution shows great promise in terms of less maintenance and significant cost savings, Biehler said. Indeed, Kirk Steudle, director of the Michigan DOT, reached out to work with a company in Japan that may now be doing an installation in Michigan. The project is a long-term project that went through various test cases in Michigan. The challenge now is to develop specifications to implement the results, Biehler said. The keys to success were to solicit ideas from internal as well as external stakeholders; to develop effective research problem statements through collaboration of Michigan DOT experts, universities, and consultants; and to involve all levels of the Michigan DOT in the development and management of the research program. Finally, as with the Arizona DOT project, a focus on implementation from the start was vital. Oregon DOT Biehler’s final example was the Oregon DOT’s research- to-application process. The process involves three tiers of staff: an overarching Research Advisory Committee (RAC), Expert Task Groups (ETGs), and staff from the Oregon DOT research section who are experts in specialty areas. The ETGs comprise staff from the Oregon DOT— frontline practitioners, managers, and researchers—and from the U.S. Federal Highway Administration. Each year, both the RAC and the ETGs set research priorities, with the RAC focusing on agency priorities and the ETGs focusing on topic area priorities. Examples of RAC priorities include improving employee safety, enhancing access and reliability, improving the environment, and reducing costs. ETG topic area priorities within the Traffic Safety and Human Factors ETG include urban–suburban design and features, continuing driver education to improve safety, and reducing the number of unsafe drivers. After the research priorities are identified, there is a three-step solicitation, evaluation, and selection process. Research staff begin the process by working with the proposer to make sure the proposal is as good as possible, namely, that it addresses the priorities and includes an implementation plan. Then, the 90 to 100 projects are assigned to their respective ETGs, which narrow them down to the two to three best proposals in the topic area, which leaves 25 proposals. Those 25 proposals are then evaluated by the final section committee, which narrows them down to the final eight to 10 projects that will receive funding.

44 t r a n s p o r t r e s e a r c h i m p l e m e n t a t i o n Common Factors of Successful DOT Projects Biehler concluded by identifying the three factors that successful DOT projects had in common: a clear description of agency priorities and the research need, a continuous focus on implementation, and joint ownership by the agency and researchers. Expediting R&D Results into Implementation Biehler also offered three ideas for how to expedite R&D results into implementation. First, he urged drafting the implementation plan at the beginning so as to carefully think through what implementation means. The implementation plan also has to identify the funding and the time frame. Either of those may change, but defining them early on provides a rigor that positions the project for success. Second, defining the pilot test (in terms of both success and failure) was helpful, as was finalizing the implementation plan at the end of R&D. Finally, having a system for tracking the research program is also helpful. Biehler stressed this last point of having a tracking program. He acknowledged that some projects may fall by the wayside, but if most of them do, the agency is wasting effort. user–procurer’s approach To The vehicle–infrasTrucTure inTeracTion Christopher Martin Christopher Martin’s presentation looked at the challenges, opportunities, and success factors for bringing complex R&D to market. For Bosch, a successful outcome of research means that someone buys it. Bosch Corporate Research Bosch is the world’s largest Tier 1 supplier of automotive products. Of 320,000 employees at Bosch, 38,500 are researchers and developers who work at 86 locations worldwide in a single network. Each working day, Bosch files an average of 16 pat- ents, which makes it one of the world’s leading compa- nies for patent applications and the one that has the most applications of any company in Germany. Bosch has invested more than €30 billion in research and develop- ment over the past 10 years. The research cuts across all sectors—transportation, energy, and consumer goods— and 80% of researchers’ time is billable, so the research function is not an overhead expense. Bosch researchers are involved in 330 different engagements, and they work with academic institutions, so research is not just internal. Although researchers keep the consumer in mind when they execute research, the research organiza- tion is foremost a technology-driven organization, not a business-driven one. Martin provided an example of Bosch’s process for complex research using the example of vehicle-to-vehicle communication [and, more broadly, V2X (i.e., beyond vehicle-to vehicle and vehicle-to-infrastructure commu- nication), because it can be vehicle-to-infrastructure, or to service providers, or to mobile phones]. Bosch is motivated to do this research because each year 5.4 mil- lion crashes occur on U.S. highways, resulting in 33,000 deaths and making traffic accidents the leading cause of death for people 4 to 34 years old. The annual cost to society of this problem is $260 billion. Besides addressing safety issues, V2X could also improve mobility issues. In the United States, the average driver is delayed more than 50 hours annually by traffic congestion. These delays amount to 2.9 billion gallons of gas wasted and an aver- age congestion cost of $80 billion annually. Automobile companies are working on V2X, as is the U.S. DOT. Together they have identified more than 100 V2X functions in the areas of safety, private-sector functions (e.g., electronic payments for services or for tolling), and public-sector functions (e.g., the optimization of traffic signal timing). These new functions have many complicated performance, safety, and security issues that involve a complex set of challenges and stakeholders, Martin said. The overall challenges are how to design the solution and develop the surrounding system to address privacy and security issues. Addressing these challenges involves Tier 1 suppliers like Bosch and its customers (the original equipment manufacturers) as well as state DOTs, telecom providers, smartphone providers, and service providers. In short, it is a large, expanded ecosystem. In addition, users have changing and varied expectations. Bosch’s R&D Process Martin next described Bosch’s R&D process, which he offered as an example of how to implement innovation in such a complex ecosystem. He prefaced his description by saying that, like ADIF, Bosch uses a Stage-Gate approach. The first step at Bosch is to scout universities and consulting agencies to see if answers already exist. If not, then Bosch will execute the study, do market research, and develop prototypes; only after that will it do project engineering. Bosch Corporate Research then engages with specific Bosch business units that have specific domain expertise. That is, Bosch Corporate Research may be expert in battery storage, but the business unit would be able to translate that knowledge specifically

45u s i n g r e s e a r c h r e s u l t s i n e f f e c t i v e w a y s into electric vehicles. In this way, the business units are involved as stakeholders. Researchers are supported by specific processes in which innovations are matched between Bosch Corporate Research and a Bosch business unit. Martin also said that Bosch has explicit transfer agreements, that is, “I will give you X, and you will do it in Way Y, and we can see if it paid off.” Project Management Bosch R&D has a project management aspect to support the innovation process, Martin said. Project managers integrate the business side and the technology side. Bosch has high expectations of its project managers and does Project Management Professional certification. Project managers are evaluated not just on hitting their milestones, but on matching up what business units want, so that research delivers the right technology for the right issue. Bosch project managers understand concepts such as the “voice of the customer,” so that research delivers what customers actually need, not just what they say they need. Project managers also do risk management, using tools such as failure mode and effects analysis, so that the innovation process is not a black box. Using such tools increases the chance of success from research to implementation, Martin said. Martin then identified the top five challenges Bosch sees in the V2X world, which include the Internet and the world of connectivity. Bosch is expert in embedded systems, which has now expanded to mean the Internet of things. That expansion challenges Bosch’s core competence. The expansion is a big challenge for Bosch and is forcing the company to change its culture. For example, what does safety mean in an Internet company? In the context of the Internet, safety is no longer just about the vehicle itself. Martin discussed the complex stakeholder environment (Figure 3) and identified the challenge of prioritizing among all these stakeholders. In addition, as end users’ expectations change (e.g., they want to bring the Internet with them into the car) and a new generation of drivers sees cars more as a burden than a right (and as something to share rather than own), Bosch has to change in response. It is an opportunity for Bosch, as a traditional embedded systems company, to engage in open, practical research and to open research test beds, partnerships, and business models that it did not have 10 years ago. Bosch is a global company, not a start-up, but it has a diversified competence base, in-house technology, and project managers who can be leveraged to rise to the challenge, Martin said. implemenTaTion of r&d resulTs in a mulTimodal urban environmenT, including public TransporT Natalia de Estevan-Ubeda Natalia de Estevan-Ubeda spoke from her experience at Transport for London (TfL), which commissions research from universities in the United Kingdom as well as from others that have transportation research labs, such as the Massachusetts Institute of Technology. Her goal for her presentation was to spark debate and pro- FIguRE 3 Challenge: Complex stakeholder environment (OEM = original equipment manufacturer). (Source: Bosch Research and Technology Center.) Standards +Certification Cost–Benefit Analysis Design + Development System Integration Security + Liability Privacy + Data Access Use Cases + Applications Field Testing Users OEMs U.S. DOT State DOTs Suppliers Challenges Stakeholders

46 t r a n s p o r t r e s e a r c h i m p l e m e n t a t i o n voke a different perspective on how to have a conversa- tion between end users (buyers) and researchers, such as through social media. TfL Context Estevan-Ubeda began by describing the context within which TfL operates. Specifically, TfL is responsible for managing all modes of transportation in London except airplanes. Roads, rail, bicycles, trams, cable cars, subways, pedestrians, barges, buses, and taxis all fall under TfL’s purview. “Our job is to keep London moving,” she said. Unlike cities such as New York City, which is laid out in straight lines, London has many winding, old roads. It also has 30 million visitors, so 30 million daily road journeys present a huge challenge. The £8 billion budget of TfL is funded by public money from the mayor of London, the government, and public transit fares. TfL must keep major highways running, as well as 6,000 sets of traffic signals, and it must integrate buses, light rail, and the subway. There is a great opportunity to make the most of all these modes and integrate transportation in a multimodal way, Estevan-Ubeda said. understanding Customer Needs Achieving success, however, requires understanding what customers want, which raises the question of who TfL’s customer is. TfL has two kinds of customers: traffic managers and travelers. When trying to get insight into what customers want, TfL discovered that the Olympic Games transformed TfL’s reputation for the better, but that reputation is not rational; that is, people do not just experience TfL’s service, they feel it. Their experience of the journey is not determined solely by reliability and journey time, but also includes the physical experience of comfort and convenience as well as an emotional experience, such as feeling safe or feeling crowded. Thus, TfL discovered from its customer research that it is not enough to provide a service, such as bus service; TfL must also support the customers’ feelings of comfort and security during their journeys. Examples of End user Involvement in Research and Implementation Next, Estevan-Ubeda provided two examples of how TfL involved end users in research and implementation. Her first example was about innovation in road space management, namely pedestrian countdown timers at traffic signals (PCaTS). These timers are not a new concept, and companies have market-ready products, but that is not enough for TfL. TfL must ensure that the products will work in the London context. Being off-the-shelf does not mean a product is finished, Estevan-Ubeda said. Indeed, TfL’s research showed that not all pedestrians understood that the green man on the traffic signal indicated an invitation to cross the street and that the black-out period showing neither the walking green man nor the stationary red man indicated that they should not cross the street. Therefore, TfL conducted off-street trials, testing a mock-up cross- ing with and without PCaTS. The research involved more than 250 pedestrians, including groups of mobility-impaired pedestrians. Questionnaires were used to establish pedestrians’ understanding and opinions of traffic signals, including PCaTS. After the off-street trials, TfL received approval to conduct on-street trials. TfL commissioned research to be done via face-to-face interviews as well as video analysis to assess pedestrian perceptions and behaviors. Sites were selected to ensure that a broad representation of pedestrians was included in the research. Results of the research showed that a clear majority of the pedestrians liked the countdown, that it reduced pedestrian uncertainty, and that it let them make more informed crossing choices. PCaTS have been introduced in London without a negative impact to safety, and the eight original trial sites have shown a 58% reduction in those who were killed or seriously injured 3 years after PCaTS introduction. In short, this example showed that TfL’s research investment is not just in the product itself but also in the application of the product. using Social Media to Inform Implementation Estevan-Ubeda’s second example was on the use of social media to inform research implementation. A year ago, TfL did not use Twitter at all, but now it uses Twitter on a daily basis to engage with customers. Customer comments on Twitter do not guide TfL’s capital program, but Twitter does provide a way to engage with customers (e.g., getting information that a traffic signal is not working or that there is a big pothole on a road). TfL has more than 210,000 followers on Twitter and uses Twitter to address reputational issues and foster engagement with the end user on the street. TfL puts out about 100 tweets a day, and 81% of its followers have changed their traffic plans as a result of TfL tweets about the congestion or delays at certain locations (Figure 4). Results of TfL’s Twitter customer satisfaction survey in 2013 showed that TfL’s real-time twitter feeds were checked often (80% of users checked them at least once a day) and that people were most likely to look at the feeds when checking Twitter generally (78%), with half

47u s i n g r e s e a r c h r e s u l t s i n e f f e c t i v e w a y s of users also referring to the feeds before they traveled or during their journey when they encountered a problem (47%). Fully 79% of TfL’s Twitter followers changed their travel plans as a result of the information provided on the feeds, and most chose a different route (59%). Applied Research TfL had to rebrand its R&D as something that was not blue-sky, because taxpayers did not want to be funding blue-sky research, Estevan-Ubeda said. Rather, TfL’s research focuses on trials and applied research. TfL research is business case driven. For there to be a business case, there must be an end user. Every research project undertaken at TfL must have a specific purpose to justify the investment. A new area in which TfL is now investing is cooperative systems and vehicle-to-infrastructure. TfL needs to know what will hit it and be at the forefront of thinking, Estevan-Ubeda said. TfL realized it did not know the implications of technologies such as intelligent transportation systems (ITS) and sensors on vehicles, so through public procurement it invited public companies, manufacturers, academics, and suppliers to engage in scenarios: “If we did this, what would it look like? What would you do first? What partners would you engage?” TfL has now commissioned some research contracts with private companies and academics. One of the obstacles that Estevan-Ubeda mentioned was that sometimes a vision is not understood in the same way by everyone. An ITS cooperative network may mean different things to different users. Similarly, the term “applied research” may have different mean- ings to different groups, so the question “what does it mean to me?” has to be addressed. TfL actively engages in budget prioritization and solves for tomorrow but not for the day after tomorrow, because TfL does not have a budget to look far forward. Closing Remarks In her closing remarks, Estevan-Ubeda offered five comments. First, she suggested involving the end user directly and indirectly (through trend data). TfL is a data-rich company and invests in translating that data into intelligence. Second, research started by the public sector is different from external innovation. Sometimes TfL will tell the market what it needs, and other times external players will tell TfL, or TfL will just let it happen. For example, in cooperative networks, TfL engaged app developers. TfL did not tell the developers want it wanted; rather, it just told them the challenge and gave them access to the data. The mandate to the app developers was to “make life easier for end users.” Thus, TfL’s influence was not directive in that case. FIguRE 4 TfL’s maximization of the use of social media. (Source: TfL.) 24/7 operations 100 tweets per day 81% of followers have changed travel plans as a result More than 210,000 followers

48 t r a n s p o r t r e s e a r c h i m p l e m e n t a t i o n Third, end users value tangible outcomes the most. Customers can see apps and the pedestrian countdowns. If TfL were to invest only in back office systems and business intelligence data fusion, customers would not see tangible outcomes. It is important for the research community to give end users some tangible benefits. Fourth, research can change opinion through facts, Estevan-Ubeda said. Finally, TfL has had success with feedback loops. TfL plans for the whole life of the project but does sense-checking to see how it applies. TfL does not want to engage in a 2-year research project and find out at the end of the 2 years that the project does not work. Rather, TfL engages in constant communication with users and industry. TfL’s Cooperative Network project, for example, began in 2012 and meets every month with both the public and private sectors. discussion Raj Rajkumar reminded participants that the purpose of this discussion was to ensure the effective use of research results at the end user level. He then proceeded to ask a question of each of the session’s presenters. First, he asked López whether his research projects are driven only by internal research activities or by external ones as well. López answered that they have two kinds of projects. The first kind has a national or international goal, and the research projects usually come from a study by commissioners of different specialties who put the problems on the table and then select the most interesting projects to pursue. Sometimes it is an internal project and other times they ask for external input. Rajkumar then asked Biehler, who was formerly secretary of the Pennsylvania DOT, to comment on how research outcomes can influence that state’s DOT planning into 2040. Biehler replied that the Pennsylvania DOT commissioned Carnegie Mellon University (CMU) to help them understand what might happen by 2040. The year 2040 was chosen because it was far enough away that the Pennsylvania DOT thought connected vehicles would be ubiquitous, and the agency wanted to know what that would mean for it. In particular, the Pennsylvania DOT wondered how it should invest and how it should operate in that changed world, because autonomous vehicles would even change driver licensing. The research was a way to stretch everyone’s thinking. Pennsylvania had recently passed a large gas tax to fund transportation expansion. However, if a $100 million transportation investment were to be made, some thought should be given to the future because traffic flow patterns will have changed in 25 years and because of the need to think multimodally. DOTs in the United States tend to think about highways first, but perhaps lanes should be set aside for autonomous vehicles or intermodal travel, Biehler said. The Pennsylvania DOT will be working in a continuous partnership with CMU researchers throughout the year-long project and meeting monthly to allow for course corrections. The research will challenge the agency to think differently about future actions as well as about workforce education. Next, Rajkumar asked Martin the following question: since Bosch both creates and consumes its own research, how are research projects spawned? Martin answered that Bosch funds its own research but also funds research by universities, with Bosch R&D staffers being charged with making the research happen. The corporate R&D staff consists of researchers who then transition the research to business units. Every 2 years, the Bosch research staff is measured on the number of $100 million projects generated. Last year, Bosch’s R&D budget was $250 million, so the return on investment (ROI) should be for six or seven $100 million projects to come out of that. The five to ten times ROI over the lifespan is aggressive, but not extremely aggressive, Martin said. The target ROI is likely to be more aggressive in the future because of the R&D options that Bosch has. Corporate R&D competes with start-ups and universities—all the different places where innovation lives—so Bosch R&D staff must deliver the most innovation bang for the research buck, even if doing so means extracting innovation from external sources like universities. Finally, Rajkumar asked Estevan-Ubeda to talk about how researchers view outcomes that affect daily operations when things go bad. Estevan-ubeda answered that the London Streets Traffic Control Center has a variety of tools that it uses for daily operations. To bring research into that Center, TfL researches software packages and has worked with operators in real time to change the way the system responds to what operators need. Through this deployment of people on the street to support engineers, research has been closely linked with daily operations. The researchers also believe that an intelligent client is a critical success factor in the implementation of research. That is, a company has to have the intelligence and know-how to operate the research, otherwise the research will not succeed. Rajkumar then opened the floor to questions from the participants. Steve Phillips asked a question about using research to inform the public with facts. Trust is an important factor in decisions to implement research, but people tend not to trust facts as much as they trust the messenger. So, Phillips asked, how do TfL and Bosch make sure that the facts they convey are trusted? Estevan-ubeda answered that TfL has a big marketing budget, which may be unusual for research but which is needed to keep TfL’s customers informed and involved in how TfL is delivering research for them.

49u s i n g r e s e a r c h r e s u l t s i n e f f e c t i v e w a y s When customers use the London subway or wait in a bus shelter, they see posters about what TfL has done. TfL influences customers in their travel choices and is responsible for telling them the status of transportation operations. Sometimes TfL is not believed. If TfL says there is no congestion someplace but customers experience a 5-minute delay, they may think that they were stuck for hours although it was only for 5 minutes. Customers see the status of the road network and where a delay is on the subway line. Marketing does not just give statistics; it also uses more appealing ways to show how research money is being invested. For example, TfL has a poster of a submarine with the TfL logo that says, “The technology that guides submarines tells you when your next bus will come.” Perception is important, Estevan- Ubeda said. If customers think there was a delay, TfL cannot tell them that there was not. Steve Andrle asked two questions of Martin: whether Bosch shares IP with university faculty and whether generating several $100 million products a year is based on forecasted sales. Martin answered that the $100 million is a measure of sales forecasts through transfer agreements with business units, so the business units work it into their sales projections. Those sales projections are then tracked by the board of directors, thereby closing the loop of tracking innovation and investing in innovation. Tracking innovation is now ingrained in the entire corporate culture and not just within R&D. A business unit says they got research from corporate R&D, and they track the resulting sales. To Andrle’s question about sharing IP versus funding university research outright, Martin explained that it varied. In some cases CMU got the IP, and in other cases the university got royalty value. Bosch has had a 25-year relationship with CMU going back to the late 1990s, when the two signed a master research agreement that defined IP rights, a gift, and sponsored research. The agreement is amended yearly. When Martin works with the University of California, Berkeley, on the other hand, the terms are different because the relationship with the university does not have a long history. Recognition of IP is bigger now, and starting from square one is hard. Most of the research is gifted research, in which one party gives X and the other keeps Y, Martin said. Peter Sweatman of the University of Michigan referred to the test beds and demonstrations mentioned in the presentations and made a plea for researchers to “get their hands dirty” and do bigger deployments. For example, he gets 40 volunteers per day who come to the University of Michigan and leave with new beneficial technology in their cars. Because the university is known as a great research institution and innovation leader, it can get a deployment of beneficial technology in 20,000 vehicles. Another point Sweatman made is that, with ubiquitous sensors, there is a great deal of data at low cost that are available to anyone, not just to transportation researchers. Greater access to data and computing power means that now anyone can do data analysis. It is not just, “This is only for professionals; do not attempt to do this at home,” Sweatman said. Researchers have to be implementers, demonstrators, and innovators. Alasdair Cain said that researchers and funders are separated because of the impartiality demanded when taxpayer dollars are used; many protections are in place by legislation. How is it possible to work closely with end users without incurring the perception of favoritism or corruption? Estevan-ubeda answered that the perception of favor- itism would be bad, so TfL makes sure that everything is done through a formal procurement process. TfL is a big entity, so the private sector wants to pitch TfL to buy X, but then TfL could be accused of provid- ing unfair advantage to one player by not talking with other players. Therefore, TfL launched open communi- cation events called notices of procurement. TfL hosts these public events, which describe what TfL is seeking to procure. Private-sector companies can register their interest and attend the events. In this way, all the com- panies receive the same message and budget figures and have the same time scale. After that event, the companies can come back and talk with TfL privately, because they might not want to share their IP at a public event. But the event is a way for companies to formally register their interest in working with TfL and then come back for one-on-one follow-up. This process provides for public engagement with everyone openly and offers a way for further follow-up privately. Biehler answered that in Pennsylvania there is a law about PPPs. Twice a year, for a period of 30 days, there is an opportunity for anyone to submit unsolicited pro- posals to the Pennsylvania DOT. There is a review board that has a rigorous process for evaluating the proposals and that then publishes the responses of its deliberations publicly. López answered that he likewise follows the laws for PPPs, so his company has to figure that separately from the rest of the company and put out a call or a contract that must be approved. Jonathan McDonald asked about the role systems engi- neering plays in the implementation of research and where the dividing line is. Martin replied that Bosch, right from the start of the project, has support processes for project management. At the start of a project, Bosch uses quality function deployment to scope out the work product and have discussions with the business units that ultimately own

50 t r a n s p o r t r e s e a r c h i m p l e m e n t a t i o n the system to which the research artifact is deployed. This process has changed over the past 25 years. The process formerly was a black box, but as the world has gotten more complex, there has been a need for more transparency so that the buyer or adopter can understand the process and the researchers, in turn, can understand the business unit’s needs. Ultimately, Bosch is a systems company—it makes the subsystems that make up a car. Bosch not only integrates the technology but integrates it from a process and project management perspective. La Torre asked whether, when expanding research to include not only the research itself but also its implementation, there was a structure for researchers to do the trials and the cost-effectiveness and feasibility studies. That is, does TfL have a research team and a separate implementation team? She remarked that in Europe they do not have implementation teams. Estevan-ubeda replied that it is a structured approach. Parts of TfL deal with the vision, scoping, and blueprint of the research; then that vision is taken up to further research (which can be internal or external). After that, TfL has systems engineering as part of its integration approach because TfL has grown to be a huge system over the years. Estevan-Ubeda offered an example of a research project done for the Oxford Circus area of London. That research was done by the teams who commissioned the research. The project had to have a business case in order for the research to get funding. Once the project was funded, it moved to the traffic infrastructure team, which did the research. Once the research was done, the project moved to the implementation team, who are experts in rolling out programs. Thus, there is a research team, but if the result is product based, then it will be deployed through the implementation team. TfL recently underwent a restructuring that resulted in separate outcomes and sponsorship teams. The sponsors generate the project, and the outcomes team ensures that the needs of the project are met. Finally, the operations team does the day-to-day running of the project. A participant asked Biehler and Martin how each organization ensures that another part of the organization is not duplicating the research. For example, another state’s DOT might be interested in knowing the solution to the problem of the elk on the road. Biehler replied that states learn from TRB and from AASHTO, and that within AASHTO there are four different regions of the United States that get together. In addition, some states find it useful to be members of the American Public Transportation Association and read its journal. Nonetheless, in some cases there is such a blur of research that it is overwhelming. People wonder which projects had good research outcomes, so ways to share research by topic are still needed. The whole connected autonomous world changes so fast, Biehler said. In short, resources from AASHTO, TRB, and the U.S. DOT are ways to learn about research that has been done. Martin answered that for Bosch it comes down to pro- cesses and people. First, the process in the six-stage chart Martin showed on one of his slides is generic compared with the actual process of doing the study, coordination, and funding. Bosch Corporate Research comprises 1,300 people around the globe. Given the way the process is coordinated and the latency of research, attempting to avoid duplication is a necessary evil and the process is not perfect. Thus, Bosch has some processes that might go slowly, but the goal is to be less wasteful. Second, the percentage of people who are researchers within Bosch Corporate Research has decreased, while the number of project managers, controllers, and evaluators has increased. The evaluators quantify the value of the research being done. Thus, the controllers are nonresearchers who support the research. Bosch also uses more social media platforms internally to let colleagues build and maintain relationships, even though they are working in different countries. For example, Bosch has a social media platform that is like an internal Twitter. The company also uses Microsoft Lync as a virtual water cooler. There is a saying at Bosch that you always meet twice—virtually and in daily relationships. Steve Andrle wondered whether, when Bosch links with other people’s systems, they find new standards. Andrle remarked that these standards could be as important as the research itself. Martin answered that yes, Bosch has a central function that engages with standards bodies and that Bosch is part of ISO 26262. Bosch has experts that participate in the standards bodies and also sends people from business units or corporate research to stay engaged. One surprise to Martin was that he never thought Bosch would be part of an Internet standards body, but now the company is participating in the Internet-related standards process of the Internet Engineering Task Force, although Bosch is not a member of it. Ángel Aparicio was curious how a big organization such as Bosch manages incremental versus radical innovative research. For example, how does Bosch decide whether to do research on the development of electric cars or research related to conventional cars? Similarly, how does TfL allocate research among the competing modes of transport? Estevan-ubeda answered that addressing one mode may have an impact on another mode, so TfL has com- peting demands and conflicting targets. If TfL does a lot of promotion of pedestrians and cycling, what is the effect on traffic flow? TfL does give attention to “managing competing demands” and has outcomes shared between all transportation modes. There are layers of meetings,

51u s i n g r e s e a r c h r e s u l t s i n e f f e c t i v e w a y s and TfL tries to avoid silos. Different boards look at cross issues and what they mean for investment and by- products, versus disruptive innovation like the submarine technology. TfL does more incremental change because it cannot afford to have something not working, but TfL can also use cooperative networks to look at more radical change further out, such as Horizon 2020. The Greater London Authority, TfL’s parent organization, is looking at Horizon 2050 for big step changes. Martin showed the innovation curve of incremental versus disruptive innovation that frames the question of incremental versus disruptive research as one of risk management (see Session 2, Figure 5). Where should a company make innovation investments? Disruption happens when a company has been focusing on X, but Y comes along, and Y is more important to customers. For example, Bosch could be focusing on reliability, which is incremental, but that reliability research could be disrupted by security. Cybersecurity has suddenly become a concern of customers because the linking of cars to the Internet makes customers vulnerable to cyber attacks or the Internet being down. Consumers now care more about security than reliability. The question is how much to invest in disruptive innovation and how to communicate the value of it. Bridging between disruptive and incremental innovation requires managing the risk portfolio, which is a big challenge. Munro asked whether Bosch purchases licenses for commercializing a technology from a third party and, if so, what percentage of licenses were third-party licenses. Martin replied that the company does purchase licenses for commercializing a technology from third parties, but that the percentage is lower because of Bosch’s large internal R&D. Kevin Womack ended the session by requesting an additional discussion point for participants to consider in their breakout groups. A report that came out in December 2013 found that the U.S. DOT portfolio was overweighted with applied research and should focus more on advanced research. On the continuum from basic research to applied research is a middle ground of advanced research, which is more like blue-sky research. Womack said that, so far, the symposium discussions around the processes and successes of applied research programs have been great, but the discussion of blue- sky research has been sparse. How does one implement the results of blue-sky research? He asked individuals in breakout groups to discuss how blue-sky research can be sold to an adopter to ensure implementation of that advanced research. If the U.S. DOT were to get only research outputs but not the valuable research outcomes of blue-sky research, the program would not last. Bridging the valley of death is hard, but he would like to see more discussion of how to get the valuable outcomes of advanced research implemented. Womack also referred back to Terry Hill’s suggestion of writing a primer on how to do implementation of advanced and applied research. He asked that, as the participants looked at the three questions in the breakout groups, they also look at those questions through the lens of advanced research. To Rajkumar’s question about whether the U.S. DOT would be creating new programs of advanced research or redirecting existing ones, Womack replied that this new focus would apply to new research programs. Hill added that he worried about where the big leaps will come from. He sits on the advisory board of Department of Engineering at the University of Cambridge, United Kingdom, and when he gets a tour of the labs and the research being done, he wonders how it will be used. What does one do with materials that are stable in two shapes? It is stunning, but what is it for? Trentacoste added that 10% ($10 million) of the U.S. Federal Highway Administration’s research budget is earmarked for advanced research. He, too, wonders how to get end users to implement the results of more exploratory research.