- Email: [email protected]
Traffic management strategies for urban networks: smart city mobility technologies
Chapter 11
Traffic management strategies for urban networks: smart city mobility technologies Alexander Skabardonis Department of Civil and Environmental Engineering, University of California, Berkeley, CA, United States
1 Existing traffic management strategies in urban networks The application of new technologies in traffic operations and management began more than 50 years ago, with the introduction of digital computers. Continuous developments in computer technology and communications have created new applications and opportunities for developing new strategies to improve mobility and safety in urban networks. This chapter focuses on strategies for arterial streets and grid networks where traffic signals are the predominant form of control. Traffic signals operate under specific timing plans (cycle length, and green times) to allocate right of way among conflicting traffic movements. Most of the existing signal systems use fixed time timing plans prepared off-line based on historical data on traffic demand. These plans are typically implemented by time of day, for example, with different plans for a.m. peak, midday, and p.m. peak periods to account for the variability of traffic. Fixed-time plans, however, cannot deal with the variability of traffic patterns throughout the day, and they become outdated because of the traffic growth and changes in traffic patterns. Several control approaches have been developed and implemented that respond to the operating conditions in the street network. Traffic-responsive and adaptive traffic signal control: These systems update the timing plans based on data from detectors located on each intersection approach and travel lane (Stevanovic, 2010). Traffic responsive systems adjust the signal settings at short time intervals based on detector data. Adaptive control policies use measured and predicted traffic data to continually optimize the signal settings. Transit signal priority (TSP): The ability of traffic signals to extend the green phase to allow an approaching transit vehicle to pass through the intersection, or reduce the duration of the active red phase so a waiting transit vehicle clears the Transportation, Land Use, and Environmental Planning. http://dx.doi.org/10.1016/B978-0-12-815167-9.00011-6 Copyright © 2020 Elsevier Inc. All rights reserved.
207
208
PART | II Strategies
intersection (Skabardonis, 2000). TSP systems consider the safety constraints for pedestrians and vehicles, and the schedule adherence and occupancy of the transit vehicle. Adaptive signal control and TSP systems have been the first examples of the active traffic management (ATM) strategies designed to manage traffic conditions in urban networks based on real-time data (FHWA, 2007). Recent computer and communications advances and the availability of new data sources provide opportunities for additional ATM strategies including but not limited to: Dynamic lane management: Changes are made to the permitted utilization of travel lanes at an intersection approach to respond to real-time traffic conditions and the associated mechanisms to communicate the changes to drivers approaching the intersection. Also, special travel lanes (e.g., continuous or intermittent bus lanes) may be dynamically designated. Queue warning systems (QWS): These systems inform drivers of downstream stop-and-go traffic caused by incidents, work zones, or other events. QWS often are coupled with speed advisory (or harmonization) systems to manage vehicle speeds to reduce the occurrence and density of shockwaves in congested traffic. Queue warning and speed harmonization can potentially reduce the number of rear-end crashes caused by stop-and-go traffic. Active parking management: These systems can provide drivers with realtime guidance to parking facilities with available capacity. They also can adjust parking prices in response to prevailing demand/availability, and activate overflow parking facilities when existing ones approach capacity. Some smart parking systems also can gather and store data on occupancy and duration by location and time of day, providing data needed for the design of time limits, if desired, development of more effective enforcement strategies, and implementation of more efficient prices. Multimodal traveler information and routing: Real-time traveler information is provided to highway users based on data from fixed detectors and mobile sources. This can include travel times and maps showing the location of construction zones, links with congestion, and location of crashes or other incidents. The severity of congestion and length of delays also can be provided. Navigation guidance has become standard and typically the user can compare available routes and modes of transport. Alternate route guidance is also provided when conditions on the primary route have deteriorated below a prescribed threshold due to congestion, incidents, or other situations. Information may be available for transit, bike, and walk modes as well as driving. For transit, travelers can use their cell phones to find transit routes and schedules from their current location or from a proposed departure point to a planned destination, and to check the arrival time of transit vehicles. The availability of timely and reliable multimodal travel information can serve as a transportation demand management (TDM) strategy that reduces the demand for roadway travel by single occupancy vehicles (ITS JPO & ITE, 2017a).
Traffic management strategies for urban networks Chapter | 11
209
Integrated corridor management (ICM): The ICM approach focuses on the coordinated management of urban networks with adjacent highway facilities (e.g., nearby parallel freeways) that comprise travel corridors (Cronin et al., 2010). Examples include coordination of arterial traffic signals with adjacent ramp meters controlling the freeway entrances to prevent queue spillovers on surface streets, and diversion of freeway traffic to arterials in case of freeway lane closures due to incidents. ICM consists of real-time operational strategies including network ATM strategies, coordination and travel information on travel modes (e.g., private auto, commuter rail, bus rapid transit), incident management, freeway control strategies, work zone planning, and inter-agency data-sharing agreements. ICM implementations employ automated decision support systems to analyze the data received and suggest appropriate control actions to the system managers.
2 Emerging applications: the promise As computer and communication technologies continue to improve, new transportation applications are being developed and tested for improving mobility and safety. These include driver assistance systems for safety, advanced detection devices, new ways for storing, processing, and analyzing data from multiple sources, and vehicle communication and automation systems. Connected Vehicles (CVs) refer to the ability of vehicles of all types (passenger cars, trucks, busses) to communicate wirelessly with other vehicles (V2V), the roadway infrastructure (V2I), such as traffic signals, as well as other non-motorized roadway users (e.g., pedestrians, bicyclists) (V2X) to support a range of safety, mobility, and environmental applications of interest to the public and private sectors (ITS JPO, 2017). Connected vehicles may bring significant additional benefits to the traffic performance of urban networks controlled by traffic signals. The information on each vehicle’s position, speed, and direction in real-time means that each vehicle becomes a sensor so continuous information on traffic demand at the vehicle’s trajectory level is provided. Furthermore, the information available on the traffic signal phase and timing (SPaT) to the vehicle through the communication of the vehicle and the infrastructure (V2I) opens opportunities for control strategies that allow vehicles to clear the intersection safely with minimum delay that also minimizes fuel use. A prototype system was developed and field demonstrated to provide in-vehicle speed advisories to drivers for minimum fuel consumption (Xia et al., 2012). The prototype system provides the recommended speed to the driver to clear the intersection without stopping based on information on the status of traffic signal (SPaT) and the vehicle’s current state. Field tests at a signalized intersection show that the proposed system achieves significant fuel savings. Furthermore, the V2X connectivity allows for developing control strategies for all road users. An example is the multimodal intelligent traffic signal system
210
PART | II Strategies
(MMITSS) (Head, 2016) that uses V2I connectivity to provide efficient control for all users at the signalized intersection—private vehicles, busses, freight vehicles, emergency vehicles, and pedestrians. The system has been successfully demonstrated in Arizona and California (Zhang et al., 2016) and is planned to be implemented at selected intersections in San Francisco under the City of San Francisco Advanced Transportation and Congestion Management Technologies Deployment Initiative (SFMTA, 2016). Emphasis in this smart city application will be placed on providing priority to transit vehicles, and dynamically extending intersection crossing times for pedestrians needing the additional time because of physical constraints. The V2V communications permits the vehicles to operate under cooperative adaptive cruise control (CACC), which allows travel at shorter headways than un-coordinated vehicles. This in turn results in higher saturation flows and shorter start-up lost times, which increase the capacity of the intersection approach and reduce delay. There is limited empirical evidence on the impacts of connected vehicles on the traffic performance. To date only a few test beds have been used on the proof of concept regarding the V2V and V2I connectivity. Currently, there are two CV pilot programs in urban networks in the United States (Tampa Florida and New York City) where a wide range of traffic applications are currently being tested (ITS JPO, 2017). In the absence of field tests, several theoretical and simulation studies have been conducted on development of control strategies using CVs and their impact assessment. Examples include queue spillback avoidance, control of congested grid networks, and dynamic lane allocation (Skabardonis, Shladover, Zhang, Zhang, & Zhou, 2013). The queue spillback avoidance strategy detects a presence of queue spillback at an intersection approach and determines the signal settings to prevent traffic from backing up the upstream intersection and avoid gridlock in grid networks. The dynamic lane grouping (DLG) control strategy changes the lane utilization at the signalized intersection approach to accommodate spatial variations in traffic demands based on real-time origin– destination information from connected vehicle data. Also, traffic safety can be improved by reducing collisions caused by vehicles running the red light, through the implementation of dynamic all-red extension (DARE) at signalized intersections (Zhang, Wang, Zhou, & Zhang, 2011). Using connected vehicle data, the vehicle trajectory through the intersection is predicted and in case there is a possibility of red light running a dynamic all red interval is inserted to prevent the potential collision. Vehicle automation (AV) is a technology actively pursued by most vehicle manufacturers and high technology companies. The concept of AV is that completely self-driven vehicles will be able to drive in all operating environments. Currently, several levels of automation have been established from level 1 (partial automation active driver engagement, lane keeping, and following distance) to level 5 (complete automation, no driver). Automated vehicles may
Traffic management strategies for urban networks Chapter | 11
211
be completely autonomous, self-driven based on safety operating rules set by the automobile manufacturers, or connected with other vehicles. Here we are concerned with the impacts of automation on the transportation system, assuming complete connectivity of automated vehicles (CAVs). Connected automated vehicles have been under research and development for many years and the technical feasibility of connected automated driving was successfully demonstrated two decades ago in “Demo97,” where a platoon of vehicles was automatically driven along a freeway section of the I-15 freeway in San Diego California, under the sponsorship of the FHWA Automated Highway Systems Consortium (TRB, 1998). However, ongoing developments in the field have led to significant differences from the earlier demonstration in terms of the technology employed, involvement of the private sector, and range of applications. Fig. 11.1 is a conceptual representation of the impacts of existing and emerging transportation technologies on mobility, safety, and air quality. The main message from this figure is the relative impacts of each technology to the performance metrics. ATM strategies offer modest improvements compared to CAVs; CAVs can drastically increase the roadway capacity (from 2000 veh/lane to 6000 veh/lane for a freeway lane) while at the same time dramatically reducing accidents (Nowakowski and Shladover, 2010). It is important to note that the benefits depend on the penetration rate of CAVs as well as the characteristics of the communication system. The capacity and safety benefits shown assume 100% penetration rate, which may take a time period of years to decades to be realized. Throughout this transition period, the transportation network will operate under a mixture of CAVs and manual vehicles. This would lower the benefits and will also require strategies for efficient management of the mixed traffic streams.
FIGURE 11.1 Potential impacts of transportation technologies.
212
PART | II Strategies
The impacts of CAVs may be far-reaching on several levels that are not captured in Fig. 11.1. They entail changes on the travel demand and behavior, supply and types of mobility services, and network design and operational performance. CAVs will enable new forms of providing mobility services. Availability of ride-share and car-sharing fleets will expand with CAVs, which can serve a wide range of users and destinations on-demand. Autonomous vehicle fleets will be able to match the specific mobility needs of individual travelers. This will significantly affect the demands on existing transportation networks, at a potential increase of vehicle miles traveled (Walker, 2017). CAVs will operate under longitudinal and lateral control, which will provide precise positioning through travel and could lead to revised design standards regarding width (narrower lanes) and other design elements of highway design. However, the implementation of narrower travel lanes for the entire roadway would require 100% penetration rate for CAVs, possibly with specialized dedicated lanes for connected, automated trucks, and busses. Alternatively, CAV lanes could be carved out of existing lanes or added along medians and shoulders. Also, CAVs can potentially travel through intersections through a “time slot” control approach that provides allocation of right of way without the need for traffic lights. The implementation of this approach is subject to several challenges, especially the accommodation of pedestrian, cyclists and other nonvehicle users.
3 Emerging applications: the implementation challenge The deployment of intelligent transportation technologies has been facing a host of barriers, only some of which are technological. Deployment costs, funding restrictions, liability concerns, uncertain demand, institutional inertia, and political challenges (Deakin, Frick, & Skabardonis, 2009). Common implementation challenges with emphasis on emerging technologies are listed in further sections.
3.1 Technology requirements The deployment of ATM strategies in urban networks requires significant infrastructure investments including sensors on each intersection approach and travel lane, and communication with the transportation management center (TMC). Pedestrian and bicycle detection systems may also be required to properly accommodate all intersection traffic. Traffic signal controllers may require upgrading to support adaptive signal control functionality. Fiber communications or other interconnects between signals may be required to support adaptive operations on an arterial corridor. CCTV cameras are helpful for location and verification of incidents and in situations in which the adaptive system is
Traffic management strategies for urban networks Chapter | 11
213
not operating properly. The implementation of dynamic facility management measures requires installation of overhead lane control signs at approximately half a mile apart. The core technology component of CAVs is wireless communications. Dedicated short range communications (DSRC) is an open-source protocol for wireless communication, intended for highly secure, high-speed wireless communication between vehicles and the infrastructure. The key functional attributes of DSRC are low latency (the delay in communications) and limited interference. Safety-related systems in the CAV environment will likely be based on DSRC. Non-safety applications may be based on different types of wireless technology (ITS JPO and ITE, 2017b). Deployment of DSRC communications in the transportation infrastructure for V2I is a significant investment and to date is available only to a limited number of government-sponsored test corridors. Looking ahead, the deployment of CAVs will generate a large amount of data. Existing traffic management centers currently do not have the capacity for such data storage, processing, and analysis. There is a need for developing a national framework that defines the process involved with the collection and sharing of data with CAVs, and creates standards, plus development of tools to support these functions (TRB, 2016).
3.2 Traffic analysis tools Existing traffic analysis tools available to transportation agencies were designed and calibrated based on data from human driven vehicles. These tools are not well suited for evaluating CAV technology applications due to their inability to incorporate vehicle connectivity as well as automation. The relationships embedded in these analysis tools may no longer hold in a connected environment. Until new tools become available, agencies need specific guidance on the application of existing tools for CAVs and their limitations. Some of the critical shortcomings of existing modeling and simulation tools include: Simulation of CAV equipment: Limitations in modeling CAV characteristics makes it difficult to evaluate the anticipated effects of various connected vehicle scenarios (e.g., equipment failures, communication latency issues, typical loads, and resource demands). This makes it difficult to assess the effectiveness of alternative CAV equipment configurations for a particular application. Currently, a fixed parameter is assumed for latency based on the communication equipment type. Limitations in modeling driver behavior: Existing simulation tools model driver behavior based on car-following, lane-changing, and queue discharge algorithms. These algorithms are currently designed to model human behavior, and have been calibrated based on vehicle trajectories of human-driven vehicles. This may be an inaccurate representation of how vehicles will operate when CV technologies and various levels of automation become prevalent (Lu, Kan, Shladover, Wei, & Ferlis, 2016). Examples include car following in
214
PART | II Strategies
mixed traffic of manual and platoon CAV vehicles, minimum headways accepted by travelers, lane changing rules, and safety margins under automated driving. Limitations in modeling driver responses to CAV applications: Several responses including driver acknowledgement, comprehension, and compliance are not currently explicitly modeled in the existing tools. Often a fixed compliance rate parameter is used to model the driver reaction to most CAV applications. However, driver response is not static; it depends on the particular application, reliability of information, trip route, and perceived position or speed improvement. Limitations in modeling CAV applications: Simulation of such applications in general is time consuming and requires custom code and logic in the model added by the analyst to support a particular application. In a recent FHWA study, it was found that 30% of the applications sponsored by FHWA in the CV pilot deployment program cannot be analyzed with the existing modeling tools (Alexiadis and Campbell, 2017).
3.3 Relationship with transportation planning studies and plans Transportation agencies conduct several studies at various time spaces (e.g., a 30 year transportation plan, a 5-year corridor investment study). Accounting in these studies for the impacts of emerging applications to a transportation network is essential but challenging. While implementation is projected, the characteristics of CAVs are still developing, and there is no empirical evidence on their impacts to the transportation system until tested. Critical questions and needs for research include:
• What values of link capacity should a transportation planner analyst assume • •
in developing a 30 year transportation and land use plan given the anticipated presence of CAVs and their penetration rate changing over time? What will be the anticipated impact and related assumptions to be made on travel behavior (mode choice, departure time choice, route choice), travel mode (vehicle ownership, new modes, shared services), and location and lifestyle choices (residential location, activities) (Peeta, 2017)? Should the transportation plan(s) be part of complete smart cities plans involving coordination of land uses, travel modes, roadway facilities and technologies? Since in the United States most states and metropolitan planning organizations have limited authority over local streets, what sorts of institutional arrangements would be needed to coordinate across jurisdictional boundaries?
3.4 Communicating the benefits of new technologies to decision-makers Elected officials and agency leaders who must make decisions about whether to invest in development and development of new technologies need clear
Traffic management strategies for urban networks Chapter | 11
215
communication on what the new technologies can do, the expected timeframe for implementation, costs involved in implementation and operations and anticipated benefits. Decision makers are not interested in the technical details of the technology. The communication regarding the benefits should be presented in a benefit-cost framework comparing the existing operations, the system operations under existing technology with optimized conditions, and operations with a the new technology. The emerging implementation of CAVs technology presents several challenges in communicating the technology to elected officials and energy leaders, given the wide availability of information from various sources on technology capabilities and impacts. Several agencies think that CAVs will mean the end of control devices that they currently are required to install and maintain (e.g., traffic signals, stop signs). Although this may be a possible scenario in the future, is not likely to be implemented in the short term. In addition, new sorts of infrastructure and equipment are likely to be needed. There is a need for clear guidance to public agencies and the regulatory requirements regarding the technology.
References Alexiadis, V., & Campbell, R. (2017). An evaluation of current simulation analysis capabilities and near-term needs for modeling connected vehicle applications. Paper presented at the 96th Annual Meeting of the Transportation Research Board, Washington, DC. Cronin, B., et al. (2010). Integrated corridor management. Public Roads, 74(3), . Deakin, E., Frick, K.T., & Skabardonis, A. (2009). Intelligent transport systems: linking technology and policy to help steer the future. Access, University of California Transportation Center, No 34. FHWA. (2007). Active traffic management: the next step to congestion management. Report FHWA-PL-07-012, Washington DC. Head, L. (2016). The multi modal intelligent traffic signal system (MMITSS): a connected vehicle dynamic mobility application. Mid Year Meeting, Traffic Signal Systems Committee, Transportation Research Board. ITS JPO, & ITE. (2017). ITS e-Primer—Module 5, Personal Transportation, USDOT, Washington, DC. Available from: https://www.pcb.its.dot.gov/eprimer.aspx. ITS JPO and ITE. (2017). ITS e-Primer—Module 14, Automated Vehicles. USDOT, Washington, DC Available from: https://www.pcb.its.dot.gov/eprimer.aspx. ITS JPO. (2017). Connected vehicles deployment program. Washington, DC: USDOT. Available from: https://www.its.dot.gov/pilots/index.htm. Lu, X., Kan, D., Shladover, S.E., Wei, D., & Ferlis, R.A. (2016). Enhanced microscopic traffic simulation model for application to connected automated vehicles. Paper presented at the 95th Annual Meeting of the Transportation Research Board. Washington, DC. Nowakowski, C., & Shladover, S.E. (2010). Cooperative adaptive cruise control: testing drivers’ choices of following distances. California PATH Research Report, UCB-ITS-PRR-2010-39, California PATH Research Report. Peeta, S. (2017). Connected and automated vehicles: modeling and management needs, FHWA workshop, getting there from here: traffic modeling, data streams and prediction for connected vehicle systems planning and operations, ISTTT22, Chicago.
216
PART | II Strategies
SFMTA. (2016). Advanced transportation and congestion management technologies deployment initiative. City of San Francisco, CA. Skabardonis, A. (2000). Control strategies for transit priority. Transportation Research Record: Journal of the Transportation Research Board, 1727(1), 20–26. Skabardonis, A., Shladover, S., Zhang, W., Zhang, L., & Zhou, K. (2013). Advanced traffic signal control algorithms. California PATH Report UCB-ITS-PRR-2013-04, prepared for the FHWA Exploratory Research Program. Stevanovic, A. (2010). Adaptive traffic control systems: domestic and foreign state of practice. NCHRP Synthesis 403, Transportation Research Board, Washington, DC. TRB. (1988). National automated highway system research program review. TRB Special Report 253, Transportation Research Board, Washington, DC. TRB. (2016). Advancing freeway operations through strategic research. Transportation Research Circular E-C218, Transportation Research Board, Washington, DC. Walker, J. (2017). The traffic jam of robots: implications of autonomous vehicles for trip-making and society. Presentation, NLBL Workshop, Designing Innovative Transportation Systems Solutions: Starting with the Data, Berkeley, CA. Xia, H., Boriboonsomsin, K., Schweizer, F., Winckler, A., Zhou, K., Zhang, W.B., & Barth, M. (2012). Field operational testing of eco-approach technology at a fixed-time signalized intersection. Proceedings of the IEEE 2012 intelligent transportation systems conference. Anchorage, AK. Zhang, L., Wang, L., Zhou, K., & Zhang, W.B. (2011). Dynamic all-red extension at signalized intersection: a framework of probabilistic modeling and performance evaluation. IEEE Transactions on ITS. Zhang, W.B., et al. (2016). Development and field testing of MMITSS in California, California PATH Research Report UCB-ITS-PRR-2016-05.
Traffic management strategies for urban networks: smart city mobility technologies Alexander Skabardonis Department of Civil and Environmental Engineering, University of California, Berkeley, CA, United States
1 Existing traffic management strategies in urban networks The application of new technologies in traffic operations and management began more than 50 years ago, with the introduction of digital computers. Continuous developments in computer technology and communications have created new applications and opportunities for developing new strategies to improve mobility and safety in urban networks. This chapter focuses on strategies for arterial streets and grid networks where traffic signals are the predominant form of control. Traffic signals operate under specific timing plans (cycle length, and green times) to allocate right of way among conflicting traffic movements. Most of the existing signal systems use fixed time timing plans prepared off-line based on historical data on traffic demand. These plans are typically implemented by time of day, for example, with different plans for a.m. peak, midday, and p.m. peak periods to account for the variability of traffic. Fixed-time plans, however, cannot deal with the variability of traffic patterns throughout the day, and they become outdated because of the traffic growth and changes in traffic patterns. Several control approaches have been developed and implemented that respond to the operating conditions in the street network. Traffic-responsive and adaptive traffic signal control: These systems update the timing plans based on data from detectors located on each intersection approach and travel lane (Stevanovic, 2010). Traffic responsive systems adjust the signal settings at short time intervals based on detector data. Adaptive control policies use measured and predicted traffic data to continually optimize the signal settings. Transit signal priority (TSP): The ability of traffic signals to extend the green phase to allow an approaching transit vehicle to pass through the intersection, or reduce the duration of the active red phase so a waiting transit vehicle clears the Transportation, Land Use, and Environmental Planning. http://dx.doi.org/10.1016/B978-0-12-815167-9.00011-6 Copyright © 2020 Elsevier Inc. All rights reserved.
207
208
PART | II Strategies
intersection (Skabardonis, 2000). TSP systems consider the safety constraints for pedestrians and vehicles, and the schedule adherence and occupancy of the transit vehicle. Adaptive signal control and TSP systems have been the first examples of the active traffic management (ATM) strategies designed to manage traffic conditions in urban networks based on real-time data (FHWA, 2007). Recent computer and communications advances and the availability of new data sources provide opportunities for additional ATM strategies including but not limited to: Dynamic lane management: Changes are made to the permitted utilization of travel lanes at an intersection approach to respond to real-time traffic conditions and the associated mechanisms to communicate the changes to drivers approaching the intersection. Also, special travel lanes (e.g., continuous or intermittent bus lanes) may be dynamically designated. Queue warning systems (QWS): These systems inform drivers of downstream stop-and-go traffic caused by incidents, work zones, or other events. QWS often are coupled with speed advisory (or harmonization) systems to manage vehicle speeds to reduce the occurrence and density of shockwaves in congested traffic. Queue warning and speed harmonization can potentially reduce the number of rear-end crashes caused by stop-and-go traffic. Active parking management: These systems can provide drivers with realtime guidance to parking facilities with available capacity. They also can adjust parking prices in response to prevailing demand/availability, and activate overflow parking facilities when existing ones approach capacity. Some smart parking systems also can gather and store data on occupancy and duration by location and time of day, providing data needed for the design of time limits, if desired, development of more effective enforcement strategies, and implementation of more efficient prices. Multimodal traveler information and routing: Real-time traveler information is provided to highway users based on data from fixed detectors and mobile sources. This can include travel times and maps showing the location of construction zones, links with congestion, and location of crashes or other incidents. The severity of congestion and length of delays also can be provided. Navigation guidance has become standard and typically the user can compare available routes and modes of transport. Alternate route guidance is also provided when conditions on the primary route have deteriorated below a prescribed threshold due to congestion, incidents, or other situations. Information may be available for transit, bike, and walk modes as well as driving. For transit, travelers can use their cell phones to find transit routes and schedules from their current location or from a proposed departure point to a planned destination, and to check the arrival time of transit vehicles. The availability of timely and reliable multimodal travel information can serve as a transportation demand management (TDM) strategy that reduces the demand for roadway travel by single occupancy vehicles (ITS JPO & ITE, 2017a).
Traffic management strategies for urban networks Chapter | 11
209
Integrated corridor management (ICM): The ICM approach focuses on the coordinated management of urban networks with adjacent highway facilities (e.g., nearby parallel freeways) that comprise travel corridors (Cronin et al., 2010). Examples include coordination of arterial traffic signals with adjacent ramp meters controlling the freeway entrances to prevent queue spillovers on surface streets, and diversion of freeway traffic to arterials in case of freeway lane closures due to incidents. ICM consists of real-time operational strategies including network ATM strategies, coordination and travel information on travel modes (e.g., private auto, commuter rail, bus rapid transit), incident management, freeway control strategies, work zone planning, and inter-agency data-sharing agreements. ICM implementations employ automated decision support systems to analyze the data received and suggest appropriate control actions to the system managers.
2 Emerging applications: the promise As computer and communication technologies continue to improve, new transportation applications are being developed and tested for improving mobility and safety. These include driver assistance systems for safety, advanced detection devices, new ways for storing, processing, and analyzing data from multiple sources, and vehicle communication and automation systems. Connected Vehicles (CVs) refer to the ability of vehicles of all types (passenger cars, trucks, busses) to communicate wirelessly with other vehicles (V2V), the roadway infrastructure (V2I), such as traffic signals, as well as other non-motorized roadway users (e.g., pedestrians, bicyclists) (V2X) to support a range of safety, mobility, and environmental applications of interest to the public and private sectors (ITS JPO, 2017). Connected vehicles may bring significant additional benefits to the traffic performance of urban networks controlled by traffic signals. The information on each vehicle’s position, speed, and direction in real-time means that each vehicle becomes a sensor so continuous information on traffic demand at the vehicle’s trajectory level is provided. Furthermore, the information available on the traffic signal phase and timing (SPaT) to the vehicle through the communication of the vehicle and the infrastructure (V2I) opens opportunities for control strategies that allow vehicles to clear the intersection safely with minimum delay that also minimizes fuel use. A prototype system was developed and field demonstrated to provide in-vehicle speed advisories to drivers for minimum fuel consumption (Xia et al., 2012). The prototype system provides the recommended speed to the driver to clear the intersection without stopping based on information on the status of traffic signal (SPaT) and the vehicle’s current state. Field tests at a signalized intersection show that the proposed system achieves significant fuel savings. Furthermore, the V2X connectivity allows for developing control strategies for all road users. An example is the multimodal intelligent traffic signal system
210
PART | II Strategies
(MMITSS) (Head, 2016) that uses V2I connectivity to provide efficient control for all users at the signalized intersection—private vehicles, busses, freight vehicles, emergency vehicles, and pedestrians. The system has been successfully demonstrated in Arizona and California (Zhang et al., 2016) and is planned to be implemented at selected intersections in San Francisco under the City of San Francisco Advanced Transportation and Congestion Management Technologies Deployment Initiative (SFMTA, 2016). Emphasis in this smart city application will be placed on providing priority to transit vehicles, and dynamically extending intersection crossing times for pedestrians needing the additional time because of physical constraints. The V2V communications permits the vehicles to operate under cooperative adaptive cruise control (CACC), which allows travel at shorter headways than un-coordinated vehicles. This in turn results in higher saturation flows and shorter start-up lost times, which increase the capacity of the intersection approach and reduce delay. There is limited empirical evidence on the impacts of connected vehicles on the traffic performance. To date only a few test beds have been used on the proof of concept regarding the V2V and V2I connectivity. Currently, there are two CV pilot programs in urban networks in the United States (Tampa Florida and New York City) where a wide range of traffic applications are currently being tested (ITS JPO, 2017). In the absence of field tests, several theoretical and simulation studies have been conducted on development of control strategies using CVs and their impact assessment. Examples include queue spillback avoidance, control of congested grid networks, and dynamic lane allocation (Skabardonis, Shladover, Zhang, Zhang, & Zhou, 2013). The queue spillback avoidance strategy detects a presence of queue spillback at an intersection approach and determines the signal settings to prevent traffic from backing up the upstream intersection and avoid gridlock in grid networks. The dynamic lane grouping (DLG) control strategy changes the lane utilization at the signalized intersection approach to accommodate spatial variations in traffic demands based on real-time origin– destination information from connected vehicle data. Also, traffic safety can be improved by reducing collisions caused by vehicles running the red light, through the implementation of dynamic all-red extension (DARE) at signalized intersections (Zhang, Wang, Zhou, & Zhang, 2011). Using connected vehicle data, the vehicle trajectory through the intersection is predicted and in case there is a possibility of red light running a dynamic all red interval is inserted to prevent the potential collision. Vehicle automation (AV) is a technology actively pursued by most vehicle manufacturers and high technology companies. The concept of AV is that completely self-driven vehicles will be able to drive in all operating environments. Currently, several levels of automation have been established from level 1 (partial automation active driver engagement, lane keeping, and following distance) to level 5 (complete automation, no driver). Automated vehicles may
Traffic management strategies for urban networks Chapter | 11
211
be completely autonomous, self-driven based on safety operating rules set by the automobile manufacturers, or connected with other vehicles. Here we are concerned with the impacts of automation on the transportation system, assuming complete connectivity of automated vehicles (CAVs). Connected automated vehicles have been under research and development for many years and the technical feasibility of connected automated driving was successfully demonstrated two decades ago in “Demo97,” where a platoon of vehicles was automatically driven along a freeway section of the I-15 freeway in San Diego California, under the sponsorship of the FHWA Automated Highway Systems Consortium (TRB, 1998). However, ongoing developments in the field have led to significant differences from the earlier demonstration in terms of the technology employed, involvement of the private sector, and range of applications. Fig. 11.1 is a conceptual representation of the impacts of existing and emerging transportation technologies on mobility, safety, and air quality. The main message from this figure is the relative impacts of each technology to the performance metrics. ATM strategies offer modest improvements compared to CAVs; CAVs can drastically increase the roadway capacity (from 2000 veh/lane to 6000 veh/lane for a freeway lane) while at the same time dramatically reducing accidents (Nowakowski and Shladover, 2010). It is important to note that the benefits depend on the penetration rate of CAVs as well as the characteristics of the communication system. The capacity and safety benefits shown assume 100% penetration rate, which may take a time period of years to decades to be realized. Throughout this transition period, the transportation network will operate under a mixture of CAVs and manual vehicles. This would lower the benefits and will also require strategies for efficient management of the mixed traffic streams.
FIGURE 11.1 Potential impacts of transportation technologies.
212
PART | II Strategies
The impacts of CAVs may be far-reaching on several levels that are not captured in Fig. 11.1. They entail changes on the travel demand and behavior, supply and types of mobility services, and network design and operational performance. CAVs will enable new forms of providing mobility services. Availability of ride-share and car-sharing fleets will expand with CAVs, which can serve a wide range of users and destinations on-demand. Autonomous vehicle fleets will be able to match the specific mobility needs of individual travelers. This will significantly affect the demands on existing transportation networks, at a potential increase of vehicle miles traveled (Walker, 2017). CAVs will operate under longitudinal and lateral control, which will provide precise positioning through travel and could lead to revised design standards regarding width (narrower lanes) and other design elements of highway design. However, the implementation of narrower travel lanes for the entire roadway would require 100% penetration rate for CAVs, possibly with specialized dedicated lanes for connected, automated trucks, and busses. Alternatively, CAV lanes could be carved out of existing lanes or added along medians and shoulders. Also, CAVs can potentially travel through intersections through a “time slot” control approach that provides allocation of right of way without the need for traffic lights. The implementation of this approach is subject to several challenges, especially the accommodation of pedestrian, cyclists and other nonvehicle users.
3 Emerging applications: the implementation challenge The deployment of intelligent transportation technologies has been facing a host of barriers, only some of which are technological. Deployment costs, funding restrictions, liability concerns, uncertain demand, institutional inertia, and political challenges (Deakin, Frick, & Skabardonis, 2009). Common implementation challenges with emphasis on emerging technologies are listed in further sections.
3.1 Technology requirements The deployment of ATM strategies in urban networks requires significant infrastructure investments including sensors on each intersection approach and travel lane, and communication with the transportation management center (TMC). Pedestrian and bicycle detection systems may also be required to properly accommodate all intersection traffic. Traffic signal controllers may require upgrading to support adaptive signal control functionality. Fiber communications or other interconnects between signals may be required to support adaptive operations on an arterial corridor. CCTV cameras are helpful for location and verification of incidents and in situations in which the adaptive system is
Traffic management strategies for urban networks Chapter | 11
213
not operating properly. The implementation of dynamic facility management measures requires installation of overhead lane control signs at approximately half a mile apart. The core technology component of CAVs is wireless communications. Dedicated short range communications (DSRC) is an open-source protocol for wireless communication, intended for highly secure, high-speed wireless communication between vehicles and the infrastructure. The key functional attributes of DSRC are low latency (the delay in communications) and limited interference. Safety-related systems in the CAV environment will likely be based on DSRC. Non-safety applications may be based on different types of wireless technology (ITS JPO and ITE, 2017b). Deployment of DSRC communications in the transportation infrastructure for V2I is a significant investment and to date is available only to a limited number of government-sponsored test corridors. Looking ahead, the deployment of CAVs will generate a large amount of data. Existing traffic management centers currently do not have the capacity for such data storage, processing, and analysis. There is a need for developing a national framework that defines the process involved with the collection and sharing of data with CAVs, and creates standards, plus development of tools to support these functions (TRB, 2016).
3.2 Traffic analysis tools Existing traffic analysis tools available to transportation agencies were designed and calibrated based on data from human driven vehicles. These tools are not well suited for evaluating CAV technology applications due to their inability to incorporate vehicle connectivity as well as automation. The relationships embedded in these analysis tools may no longer hold in a connected environment. Until new tools become available, agencies need specific guidance on the application of existing tools for CAVs and their limitations. Some of the critical shortcomings of existing modeling and simulation tools include: Simulation of CAV equipment: Limitations in modeling CAV characteristics makes it difficult to evaluate the anticipated effects of various connected vehicle scenarios (e.g., equipment failures, communication latency issues, typical loads, and resource demands). This makes it difficult to assess the effectiveness of alternative CAV equipment configurations for a particular application. Currently, a fixed parameter is assumed for latency based on the communication equipment type. Limitations in modeling driver behavior: Existing simulation tools model driver behavior based on car-following, lane-changing, and queue discharge algorithms. These algorithms are currently designed to model human behavior, and have been calibrated based on vehicle trajectories of human-driven vehicles. This may be an inaccurate representation of how vehicles will operate when CV technologies and various levels of automation become prevalent (Lu, Kan, Shladover, Wei, & Ferlis, 2016). Examples include car following in
214
PART | II Strategies
mixed traffic of manual and platoon CAV vehicles, minimum headways accepted by travelers, lane changing rules, and safety margins under automated driving. Limitations in modeling driver responses to CAV applications: Several responses including driver acknowledgement, comprehension, and compliance are not currently explicitly modeled in the existing tools. Often a fixed compliance rate parameter is used to model the driver reaction to most CAV applications. However, driver response is not static; it depends on the particular application, reliability of information, trip route, and perceived position or speed improvement. Limitations in modeling CAV applications: Simulation of such applications in general is time consuming and requires custom code and logic in the model added by the analyst to support a particular application. In a recent FHWA study, it was found that 30% of the applications sponsored by FHWA in the CV pilot deployment program cannot be analyzed with the existing modeling tools (Alexiadis and Campbell, 2017).
3.3 Relationship with transportation planning studies and plans Transportation agencies conduct several studies at various time spaces (e.g., a 30 year transportation plan, a 5-year corridor investment study). Accounting in these studies for the impacts of emerging applications to a transportation network is essential but challenging. While implementation is projected, the characteristics of CAVs are still developing, and there is no empirical evidence on their impacts to the transportation system until tested. Critical questions and needs for research include:
• What values of link capacity should a transportation planner analyst assume • •
in developing a 30 year transportation and land use plan given the anticipated presence of CAVs and their penetration rate changing over time? What will be the anticipated impact and related assumptions to be made on travel behavior (mode choice, departure time choice, route choice), travel mode (vehicle ownership, new modes, shared services), and location and lifestyle choices (residential location, activities) (Peeta, 2017)? Should the transportation plan(s) be part of complete smart cities plans involving coordination of land uses, travel modes, roadway facilities and technologies? Since in the United States most states and metropolitan planning organizations have limited authority over local streets, what sorts of institutional arrangements would be needed to coordinate across jurisdictional boundaries?
3.4 Communicating the benefits of new technologies to decision-makers Elected officials and agency leaders who must make decisions about whether to invest in development and development of new technologies need clear
Traffic management strategies for urban networks Chapter | 11
215
communication on what the new technologies can do, the expected timeframe for implementation, costs involved in implementation and operations and anticipated benefits. Decision makers are not interested in the technical details of the technology. The communication regarding the benefits should be presented in a benefit-cost framework comparing the existing operations, the system operations under existing technology with optimized conditions, and operations with a the new technology. The emerging implementation of CAVs technology presents several challenges in communicating the technology to elected officials and energy leaders, given the wide availability of information from various sources on technology capabilities and impacts. Several agencies think that CAVs will mean the end of control devices that they currently are required to install and maintain (e.g., traffic signals, stop signs). Although this may be a possible scenario in the future, is not likely to be implemented in the short term. In addition, new sorts of infrastructure and equipment are likely to be needed. There is a need for clear guidance to public agencies and the regulatory requirements regarding the technology.
References Alexiadis, V., & Campbell, R. (2017). An evaluation of current simulation analysis capabilities and near-term needs for modeling connected vehicle applications. Paper presented at the 96th Annual Meeting of the Transportation Research Board, Washington, DC. Cronin, B., et al. (2010). Integrated corridor management. Public Roads, 74(3), . Deakin, E., Frick, K.T., & Skabardonis, A. (2009). Intelligent transport systems: linking technology and policy to help steer the future. Access, University of California Transportation Center, No 34. FHWA. (2007). Active traffic management: the next step to congestion management. Report FHWA-PL-07-012, Washington DC. Head, L. (2016). The multi modal intelligent traffic signal system (MMITSS): a connected vehicle dynamic mobility application. Mid Year Meeting, Traffic Signal Systems Committee, Transportation Research Board. ITS JPO, & ITE. (2017). ITS e-Primer—Module 5, Personal Transportation, USDOT, Washington, DC. Available from: https://www.pcb.its.dot.gov/eprimer.aspx. ITS JPO and ITE. (2017). ITS e-Primer—Module 14, Automated Vehicles. USDOT, Washington, DC Available from: https://www.pcb.its.dot.gov/eprimer.aspx. ITS JPO. (2017). Connected vehicles deployment program. Washington, DC: USDOT. Available from: https://www.its.dot.gov/pilots/index.htm. Lu, X., Kan, D., Shladover, S.E., Wei, D., & Ferlis, R.A. (2016). Enhanced microscopic traffic simulation model for application to connected automated vehicles. Paper presented at the 95th Annual Meeting of the Transportation Research Board. Washington, DC. Nowakowski, C., & Shladover, S.E. (2010). Cooperative adaptive cruise control: testing drivers’ choices of following distances. California PATH Research Report, UCB-ITS-PRR-2010-39, California PATH Research Report. Peeta, S. (2017). Connected and automated vehicles: modeling and management needs, FHWA workshop, getting there from here: traffic modeling, data streams and prediction for connected vehicle systems planning and operations, ISTTT22, Chicago.
216
PART | II Strategies
SFMTA. (2016). Advanced transportation and congestion management technologies deployment initiative. City of San Francisco, CA. Skabardonis, A. (2000). Control strategies for transit priority. Transportation Research Record: Journal of the Transportation Research Board, 1727(1), 20–26. Skabardonis, A., Shladover, S., Zhang, W., Zhang, L., & Zhou, K. (2013). Advanced traffic signal control algorithms. California PATH Report UCB-ITS-PRR-2013-04, prepared for the FHWA Exploratory Research Program. Stevanovic, A. (2010). Adaptive traffic control systems: domestic and foreign state of practice. NCHRP Synthesis 403, Transportation Research Board, Washington, DC. TRB. (1988). National automated highway system research program review. TRB Special Report 253, Transportation Research Board, Washington, DC. TRB. (2016). Advancing freeway operations through strategic research. Transportation Research Circular E-C218, Transportation Research Board, Washington, DC. Walker, J. (2017). The traffic jam of robots: implications of autonomous vehicles for trip-making and society. Presentation, NLBL Workshop, Designing Innovative Transportation Systems Solutions: Starting with the Data, Berkeley, CA. Xia, H., Boriboonsomsin, K., Schweizer, F., Winckler, A., Zhou, K., Zhang, W.B., & Barth, M. (2012). Field operational testing of eco-approach technology at a fixed-time signalized intersection. Proceedings of the IEEE 2012 intelligent transportation systems conference. Anchorage, AK. Zhang, L., Wang, L., Zhou, K., & Zhang, W.B. (2011). Dynamic all-red extension at signalized intersection: a framework of probabilistic modeling and performance evaluation. IEEE Transactions on ITS. Zhang, W.B., et al. (2016). Development and field testing of MMITSS in California, California PATH Research Report UCB-ITS-PRR-2016-05.