Field implementation and testing of an automated eco-cooperative adaptive cruise control system in the vicinity of signalized intersections
Introduction
The transportation sector is the second largest contributor to an increase in fuel consumption and emission levels in the United States (Agency, 2010). A rise in vehicle travel miles and, consequently, traffic congestion has led to greater fuel consumption and pollution. One solution, the concept of eco-driving, has been studied widely as a way to mitigate the environmental impacts of transportation (Rafael, 2006).
Eco-driving is a driving style that is considered to be both economical and ecologically beneficial. It has been defined by several researchers as the smooth driving resulting from adjustments in driver behavior (De Vlieger et al., 2000). Consequently, researchers have provided tips for drivers (Ando et al., 2010) on how to effectively minimize fuel consumption and emission levels. These tips are:
- •
Avoid hard braking, sudden acceleration/deceleration, and idling.
- •
Limit air conditioner use.
- •
Do not warm engine.
- •
Upshift early.
- •
Maintain steady speed.
- •
Look ahead while driving.
- •
Limit starting speed in the first 5 s to 12.4 mi/h (20 km/h).
In addition, recent research has shown how onboard driver assistance systems can be installed in vehicles to help drivers successfully follow eco-driving principles (Ando et al., 2010, Vagg, 2013). These devices can deliver “static” advice to drivers such as “accelerate slowly,” “avoid hard braking,” and “reduce high speed.” However, the impact of static advice on fuel consumption and emission levels is limited.
By taking advantage of real-time traffic sensing and infrastructure information, onboard driver assistance systems can deliver “dynamic” advice to drivers that can help them drive smoothly and efficiently. In particular, establishing communication between the vehicle and the traffic signal controller is a powerful tool to make driver behavior smoother, thus reducing vehicle fuel consumption levels. Various messages can be exchanged between the vehicle and the traffic signal controller to provide better advice to drivers. Among these messages is the Signal Phasing and Timing (SPaT) message, which can provide approaching vehicles with valuable advance information about upcoming traffic signal timing changes. This information could help drivers make early decisions about whether to proceed or stop safely and smoothly.
With the help of vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication, algorithms are being developed that utilize SPaT data together with queue information to optimize vehicle trajectories in the vicinity of signalized intersections (Mandava et al., 2009, Asadi and Vahidi, 2011, Barth, et al., 2011, Sun, 2013). In other words, vehicles can communicate with traffic signal controllers, receive SPaT information, and adjust their speeds accordingly and automatically. Several Eco-Speed control (ESC) algorithms have been introduced to provide an optimal trajectory through signalized intersections. However, none explicitly attempt to minimize vehicle fuel consumption levels. In addition, most of the proposed algorithms have been developed and tested in traffic simulation environments where recommended speeds are enforced and many issues, such as the delay in the system and human-vehicle interaction, are not considered.
Kamalanathsharma (2014) developed a new ESC algorithm named Eco-Cooperative Adaptive Cruise Control (Eco-CACC). The Eco-CACC system computes and recommends real-time, fuel-efficient speeds using V2V- and V2I-communicated data within the vicinity of the traffic signalized intersection. The objective function of the Eco-CACC system is the explicit minimization of the total fuel consumed to travel from some distance upstream of the intersection to a distance downstream of the intersection. The Eco-CACC system was tested and evaluated in a simulation environment. A reduction in fuel consumption of over 30% was achieved. In a preliminary field test, we addressed the implementation issues and challenges of the field application of the Eco-CACC system (Chen, 2016) and showed that the Eco-CACC system is practical, applicable and very promising in terms of fuel consumption and travel time savings. Also, In Almannaa (2017), we tested the Eco-CACC system (with only human-vehicle interaction) in a real environment and found it could save up to 19% of fuel consumption within the vicinity of signalized intersections. Furthermore, we studied the queue effects, multiple intersections, and interaction of Eco-CACC and non-Eco-CACC vehicles in a simulation environment (Yang et al., 2017, Yang and Almutairi, 2017).
In this paper, we extend our work by conducting a more extensive controlled field experiment using 32 participants on the Virginia Smart Road test facility at the Virginia Tech Transportation Institute (VTTI) with the goal of evaluating the Eco-CACC system with regard to fuel consumption and travel time savings. This experiment is mainly testing the human-vehicle interaction (driver perception/reaction delay) under controlled circumstances as well as considering other technical issues: communication latency, system malfunction, and data collection error. Three different scenarios were considered: normal driving, driving with recommended speed information provided to drivers (a manual Eco-CACC system), and finally automated driving (an automated Eco-CACC system). Previous studies have shown that visual messages are less efficient and more distracting for drivers compared to audio messages (Jamson et al., 2015, Horberry, 2006, Tang, 2016). Consequently, the computed speeds in the manual Eco-CACC system were delivered to drivers as audio messages.
The following sections provide a brief literature review; a brief description of the algorithm used; the experimental design used to gather the data; field data analysis and findings; and finally the conclusions of the research and recommendations for future work.
Section snippets
Background
Researchers have recently realized the significant increase in vehicle fuel consumption levels that can be attained in the vicinity of signalized intersections (Courage and Parapar, 1975). When coming close to a signalized intersection, drivers are unaware of exactly when the traffic signal indication will change. Consequently, drivers may have to accelerate/decelerate aggressively to respond to these changes. This results in non-smooth driving behavior. Non-smooth driving (e.g., hard barking,
Methodology
The control algorithm in the Eco-CACC system was developed in the previous work (Chen, 2016, Chen, et al., 2017, Rakha, et al., 2016) to provide a fuel-optimized speed profile both upstream and downstream of a signalized intersection. The objective function of the proposed algorithm is the explicit minimization of the total fuel consumed to travel from some distance upstream of the intersection to a distance downstream of the intersection. In addition, various constraints are constructed using
Field data analysis
Vehicle trip information (e.g., the current and recommended speed, distance to stop line, time, and SPaT information) was recorded each decisecond. Only the data in the Eco-CACC range (820 feet [250 m] upstream to 590 feet [180 m] downstream) were extracted and analyzed. In total, 1536 trips were recorded for 32 participants.
The in-field data results show that the fuel consumption measurements were not reliable as they do not capture the sudden changes in speeds and, also, give a value of zero
Experiment
The results show that both the manual and automated Eco-CACC systems always improved the vehicle trajectory for all treatment combinations. In particular, they decreased acceleration and deceleration maneuvers and provided a smooth speed profile. This can be seen in Fig. 11, in which an example vehicle speed profile and trajectory are compared for the three scenarios for each red indication offset value. Note that zero marks time when the test vehicle enters the range of the Eco-CACC system. In
Conclusions and recommendations for future research
The research presented in this paper developed, field implemented, field tested, and quantified the benefits of an Eco-CACC system that computes a vehicle’s fuel-optimal trajectory in real-time using SPaT information received from a downstream traffic signal controller. A controlled field experiment involving 32 participants was conducted on the Smart Road test facility at VTTI. The experiment included three scenarios: normal driving, driving with the manual Eco-CACC system enabled, and driving
Acknowledgements
This research effort was jointly funded by the TranLIVE University Transportation Center, NPRP Grant # 5-1272-1-214 from the Qatar National Research Fund (a member of The Qatar Foundation), and the University Equity and Mobility Center (UMEC). The authors acknowledge the help of researchers from the hardware group at VTTI for their assistance in developing the hardware and software environment.
References (53)
- et al.
Environmental effects of driving behaviour and congestion related to passenger cars
Atmos. Environ.
(2000) - et al.
Drivers’ ability to learn eco-driving skills; effects on fuel efficient and safe driving behaviour
Transport. Res. Part C: Emerg. Technol.
(2015) Driver distraction: the effects of concurrent in-vehicle tasks, road environment complexity and age on driving performance
Accid. Anal. Prev.
(2006)- et al.
A fuel economy optimization system with applications in vehicles with human drivers and autonomous vehicles
Transport. Res. Part D: Transport Environ.
(2011) A trajectory smoothing method at signalized intersection based on individualized variable speed limits with location optimization
Transport. Res. Part D: Transport Environ.
(2018)- et al.
Comparing effects of eco-driving training and simple advices on driving behavior
Procedia – Social Behav. Sci.
(2012) Using on-board logging devices to study the longer-term impact of an eco-driving course
Transport. Res. Part D: Transport Environ.
(2009)- et al.
Energy and emissions impacts of a freeway-based dynamic eco-driving system
Transport. Res. Part D: Transport Environ.
(2009) Virginia tech comprehensive power-based fuel consumption model: model development and testing
Transport. Res. Part D: Transport Environ.
(2011)Fuel economy impacts of manual, conventional cruise control, and predictive eco-cruise control driving
Int. J. Transp. Sci. Technol.
(2013)