0
Select Articles

Vehicle and Powertrain Optimization for Autonomous and Connected Vehicles OPEN ACCESS

[+] Author Notes
Yunli Shao, Mohd Azrin Mohd Zulkefli, Zongxuan Sun

Department of Mechanical Engineering University of Minnesota

Yunli Shao received the B.S. degree in Mechanical Engineering from Shanghai Jiao Tong University, Shanghai, China, in 2013, and the M.S. degree in Mechanical Engineering from University of Michigan, Ann Arbor, in 2015. He is currently working toward the Ph.D. degree at the Automotive Propulsion Control Lab, Mechanical Engineering, University of Minnesota, Twin Cities. His research interests are the control, optimization and evaluation of connected vehicles.

Mohd Azrin Mohd Zulkefli received the B.S, M.Sc. and Ph.D. degrees in Mechanical Engineering from the University of Minnesota, Twin Cities, in 2006, 2009 and 2017, respectively. He is currently a postdoctoral researcher with the Automotive Propulsion Control Lab, Mechanical Engineering, University of Minnesota, Twin Cities. His research interests are optimization and evaluation of hybrid vehicles in the connected vehicle setting.

Zongxuan Sun received the B.S. degree in automatic control from Southeast University, Nanjing, China, in 1995, and the M.S. and Ph.D. degrees in mechanical engineering from the University of Illinois at Urbana─Champaign, Champaign, IL, USA, in 1998 and 2000, respectively. He is currently Professor of Mechanical Engineering at the University of Minnesota, Minneapolis, MN, USA, and the Co-Deputy Director of the NSF Engineering Research Center of Compact and Efficient Fluid Power. He was a Staff Researcher (2006─2007) and a Senior Researcher (2000─2006) at General Motors Research and Development Center, Warren, MI, USA. He has published more than 120 refereed technical papers and received 20 U.S. patents. His research interests include controls and mechatronics with applications to automotive propulsion systems.

Mechanical Engineering 139(09), S19-S23 (Sep 01, 2017) (5 pages) Paper No: ME-17-SEP6; doi: 10.1115/1.2017-Sep-6

This article discusses the potential of using autonomous and connected vehicle (CV) technologies to save energy. It also focuses on the potential energy savings of internal combustion engine-based vehicles (ICVs) and hybrid electric vehicles (HEVs). An example of vehicle and powertrain co-optimization for HEV eco-approaching and departure is also given. CV technologies are gaining increasing attention around the world. Vehicle-to-vehicle (V2V) communication and vehicle-to-infrastructure (V2I) communication enable real-time access to traffic information that was not available before, including preceding vehicles’ location, speed, pedal position, traffic signal phasing and timing (SPaT). The example shown in this article demonstrates the potential benefits from vehicle and powertrain co-optimization by investigating an eco-approaching and departure application. More research in this area can offer more mature solutions to implement such optimization in a real-production vehicle.

Transportation accounts for about 30 percent of total greenhouse gas emissions in the United States [1]. There have been increasing concerns about its impact on the environment and human health. Reducing emissions and improving the fuel efficiency of vehicles have become challenging and urgent issues. This article will discuss the potential of using autonomous and connected vehicle (CV) technologies to save energy.

Recently, CV technologies are gaining increasing attention around the world. Vehicle-to-vehicle (V2V) communication and vehicle-to-infrastructure (V2I) communication enable real-time access to traffic information that was not available before, including preceding vehicles’ location, speed, pedal position, traffic signal phasing and timing (SPaT) and the like. This newly available information allows vehicles to anticipate future driving conditions and vehicle load changes. Many of the connected vehicle applications (CVAs) are targeting safety concerns of transportation, but recently energy savings through CVAs have also attracted more attention. There are several emerging CVAs for energy savings including eco-cooperative adaptive cruise control (Eco-CACC), eco-approaching and departure (Eco-AD), and ecorouting. Fuel efficiency can be improved at two levels: vehicle level and powertrain level. By adjusting vehicle speed and the following distance, vehicle power demand and fuel consumption can be optimized. For the same power demand, powertrain operation can be optimized to improve fuel efficiency. Co-optimization of the vehicle dynamics and powertrain operation offers the maximum potential for energy savings.

The remainder of this article is organized as follows: First, energy saving potentials from vehicle dynamics and powertrain optimization for CVAs are explained. Then a co-optimization framework is described. Finally, a cooptimized Eco-AD application is shown as an example.

From 1980 to 2014, the average mile per gallon (MPG) of a light-duty vehicle has increased by about 50 percent [2]. Vehicles have become more efficient with many technologies that improve aerodynamic design, reduce vehicle weight, and increase powertrain efficiency (e.g., variable valve timing, turbochargers and superchargers). However, the way the vehicle is driven can also impact fuel efficiency. A human driver reacts to inputs of the surrounding vehicles, road and traffic conditions. The commands to the vehicle (pedal, brake and steering) are updated after the driver has seen changes in the driving situation. This reaction process can result in driving behaviors that are not fuel efficient. For example, the driver has to brake hard near the intersection when the signal suddenly changes to red. During braking, energy is significantly wasted since most of the vehicle kinetic energy is dissipated.

With CV technologies, rich information can be transmitted to improve the control decisions of the vehicle. The vehicle can have not only inputs from the immediate preceding vehicle, but also inputs from other vehicles in the platoon (such as the leading vehicle) and information of the future traffic signals. The driver can anticipate the upcoming changes in the driving situation and then avoid hard braking and idling. In addition, with the help of automation, the vehicle can accelerate more smoothly to avoid abrupt accelerations. Most of the time, abrupt acceleration is unnecessary and can significantly increase the instantaneous power demand of the vehicle and thus increase fuel consumption.

The most common passenger vehicles today are internal combustion engine-based vehicles (ICVs) and hybrid electric vehicles (HEVs). In the near future, these vehicle types are still likely to be dominant in the market. Therefore, in this section, we will mainly focus on the potential energy savings of ICVs and HEVs.

Internal Combustion Engine-Based Vehicles

Usually, the most efficient operating region of an internal combustion engine (ICE) is the region near high torque and low speed, as shown in Figure 1. The efficiency of the engine is reflected by the brake-specific fuel consumption (BSFC) contours in the figure. It measures how many grams of fuel the engine will consume per hour to generate one unit of power. So a smaller BSFC means a more efficient operating condition. In the ideal case, the engine should always operate in the peak efficiency region. However, the engine is sized for the most demanding performance criterion. Under an average driving scenario, the engine will only operate on partial load conditions and it may not be able to operate in the most efficient region. With the transmission, engine operating conditions can be changed according to different vehicle operating conditions. For most common ICVs, an automatic gear transmission is used with a fixed gear shift schedule to determine the speed ratio between the engine output and the wheels of the vehicle. However, this shift schedule is again calibrated to satisfy the most demanding performance requirement. To ensure the vehicle has good drivability under all driving scenarios, the engine cannot always operate at peak efficiency.

FIGURE 1 Brake-specific fuel consumption (BSFC) of an internal combustion engine.

Grahic Jump LocationFIGURE 1 Brake-specific fuel consumption (BSFC) of an internal combustion engine.

So how can connectivity help? Connectivity makes it possible to bring an extra degree of freedom by adapting the gear shift schedule based on real-time vehicle speed and power demand. Fundamentally, the powertrain efficiency is limited by the challenge that the vehicle speed and power demand are varying in real-time and the engine efficiency depends on speed and load conditions. With information from V2V and V2I, the powertrain controller can anticipate the upcoming changes in the power demand and decide whether it should be tailored for fuel economy or performance (drivability). For example, if the powertrain controller predicts that the driver will be cruising in the future, it can shift to higher gears earlier and then operate the engine in more efficient regions.

Hybrid Electric Vehicles

The powertrain architecture of an HEV consists of the conventional ICE and an alternative electrical power source. Using the extra degree of freedom enabled by the alternative electrical power source, the engine operating condition can be optimized independently from the vehicle power demand. To fully realize the fuel saving potentials of HEVs, the power management controller of the hybrid powertrain must be carefully designed. This controller determines the proper power-split between the two power sources, the ICE and the electrical power source.

The current HEVs in the market usually employ heuristic or ad-hoc power management strategies that may not be optimal under a real driving scenario. These strategies use the electrical power source conservatively to prevent depleting the battery. The powertrain optimization is achieved by determining the best power-split ratio in real-time. The optimal power management can be achieved if the future vehicle speed trajectory can be anticipated a priori. This is infeasible without real-time traffic information from the connectivity. With this information, the optimal controller can anticipate the upcoming vehicle trajectories accurately and then select the appropriate power source intelligently. For example, when approaching an intersection, if the vehicle knows the traffic signal will turn red and a stop is necessary, the controller can choose to use more electrical power to drive the vehicle approaching the intersection. This way, the engine will work less, hence use less fuel, while the battery will be replenished through regenerative braking when the vehicle decelerates to a stop.

The previous sections have explained how connectivity can help save energy from vehicle dynamics and powertrain operation optimization. Now the question is how to design an intelligent controller to improve overall fuel economy. Such controller can be obtained by solving a mathematical optimization problem. This co-optimization of both vehicle and powertrain requires connectivity to anticipate upcoming traffic information and partial or full vehicle automation to execute the optimal control law on the vehicle. An example of a threestep framework includes:

First, model the longitudinal vehicle dynamics and the powertrain dynamics. The optimization objective is to minimize the fuel consumption. The optimization variables (control means) are the desired vehicle acceleration, engine throttle, gear shift scheduling, and hybrid power usage (if it exists). The traffic information (signal phase and timing (SPaT), preceding vehicles’ speed and acceleration, dynamic speed limit, etc.) is the input.

Second, define the energy saving optimization problem and set the constraints according to the specific CVAs. For example, in Eco-AD, the vehicle should maintain a safe following distance, obey the speed limit and pass the intersection only when the signal is green.

Third, develop real-time implementable solutions to the above optimization problem. Potential methods include both the indirect method, which is based on the calculus of variation and Pontryagin's minimum principle (PMP), and the direct method, which is based on discretizing the problem into a nonlinear programming (NLP) problem and solving it using numerical optimization tools.

There are several challenges for developing and implementing a co-optimization controller:

  1. Real-time computational burden. The computational time depends on the solving methods as well as the complexity of the models and constraints. A complex model can reflect the vehicle and powertrain dynamics more accurately but at the cost of increasing computational effort. On the other hand, an over-simplified model may lead to a controller that is not truly optimal due to the modeling errors. The co-optimization problem should be formulated such that it achieves a balance between performance and computational effort.

  2. Reliability and accuracy of traffic information and prediction. A traffic prediction is necessary so as to estimate preceding vehicles’ future positions and speeds. The effectiveness of the co-optimization algorithm relies on the accuracy of the prediction. Developing a reliable traffic prediction is crucial. In addition, the optimal controller should also be robust to ensure energy savings and safe operation even under uncertainties.

  3. Interactions between vehicle automation and driver. When there is vehicle automation, the engaging and disengaging of drivers to control the vehicle can be complex. The interactions between the optimal controller and the driver need to be carefully designed to ensure the driver understands what driving functions are handled by the automation.

  4. Availability of control means. To implement the co-optimization controller, real-time access to the vehicle and powertrain control (throttle, gear shift, hybrid power usage, etc.) is necessary. Access to these control means may not be directly available. Expertise is required during implementation.

In this section, an example of vehicle and powertrain co-optimization for HEV eco-approaching and departure is given. The target vehicle follows a preceding vehicle to pass an intersection and the performance of the two vehicles is evaluated in simulation. For simplicity, it is assumed that the target vehicle perfectly knows the future signal phase and timing, and the preceding vehicle's trajectory. The trajectory of the preceding vehicle was recorded from an actual vehicle driven by a human driver during a field test on Trunk Highway 55 in Minnesota. The target vehicle optimizes both the vehicle speed and powertrain power-split as it follows the preceding vehicle. The co-optimization framework is based on [3]. The two-level optimization is performed in a consecutive order. First, the vehicle dynamics controller solves an optimal vehicle acceleration trajectory by using the preceding vehicle's trajectory and the signal information. Next, the optimal acceleration profile is sent to the powertrain controller to optimize engine operating points and the power-split simultaneously. The details of the powertrain optimization can be found in [4].

Simulation results are shown in Figure 2 and Figure 3. In Figure 2, both vehicles are using a rule based (heuristic) powertrain controller while the target vehicle's speed is optimized. In Figure 3, the preceding vehicle uses an optimal powertrain controller while the target vehicle's speed and powertrain operation are cooptimized. Figure 2a and Figure 2b show that, at first, without traffic signal information, the preceding vehicle accelerates and then decelerates when the vehicle is within a comfortable stopping distance to the red traffic signal. In contrast, the target vehicle knows when the signal will turn green, and therefore maintains its speed. It then decelerates slowly, while maintaining an appropriate following distance, before accelerating when the signal turns green. Without knowing future signal information, the preceding vehicle's performance is constrained based on current information, which forces it to be reactive and decelerate more when the signal is red. The target vehicle accelerates and decelerates less than the preceding vehicle, hence requiring less power as reflected by the lower engine and battery power (Figure 2f and Figure 2g).

FIGURE 2 Vehicles and powertrain dynamics with rule-based powertrain controller.

Grahic Jump LocationFIGURE 2 Vehicles and powertrain dynamics with rule-based powertrain controller.

FIGURE 3 Vehicles and powertrain dynamics with optimal powertrain controller.

Grahic Jump LocationFIGURE 3 Vehicles and powertrain dynamics with optimal powertrain controller.

The effects of powertrain optimization can be seen by comparing Figure 2 and Figure 3. As observed, with powertrain operation optimization, the engine speeds in Figure 3d are generally lower than those in Figure 2d. In addition, engine idling is also either eliminated or minimized, as can be seen from Figure 2c and Figure 3c. These correspond to more efficient engine operating regions, which result in better fuel economy for both vehicles (Figure 3e) compared to vehicle dynamics optimization only (Figure 2e). Moreover, Figure 2h and Figure 3h show that for both vehicles, at the beginning of the cycle, the optimal powertrain controller uses more battery to propel the vehicle since it anticipates a future braking event that will replenish the battery charge. Figure 4 shows a comparison of fuel consumption and fuel benefits. With vehicle dynamics and powertrain operation co-optimization, the target vehicle fuel saving is about 23 percent compared to the preceding vehicle. The vehicle dynamics optimization brings the target vehicle about 17 percent fuel saving. Furthermore, with powertrain operation optimization, the preceding vehicle can also achieve about 10 percent fuel improvement.

FIGURE 4 Comparison of fuel consumption and fuel benefits.

Grahic Jump LocationFIGURE 4 Comparison of fuel consumption and fuel benefits.

With the enabling technology of connected vehicles, there are opportunities to save energy through both vehicle dynamics and powertrain operation optimization. The vehicle can be driven more proactively to anticipate the upcoming traffic situations and the powertrain can correspondingly operate in the most efficient regions. The example shown in this article demonstrates the potential benefits from vehicle and powertrain co-optimization by investigating an eco-approaching and departure application. Further research in this area can offer more mature solutions to implement such optimization in a real production vehicle.

United States Environmental Protection Agency, “Sources of Greenhouse Gas Emissions”, https://www.epa.gov/ghgemissions/sources-greenhouse-gas-emissions, accessed June 10, 2017.
United States Department of Transportation, “Average Fuel Efficiency of U.S. Light Duty Vehicles”, https://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_04_23.html, accessed June 10, 2017.
Hu, J., Shao, Y., Sun, Z., Wang, M., Bared, J. and Huang, P., “Integrated Optimal Eco-Driving On Rolling Terrain for Hybrid Electric Vehicle with Vehicle-Infrastructure Communication”, Transportation Research Part C, 68, 2016, pp. 228– 244. [CrossRef]
Mohd Zulkefli, M.A., “Connected Hybrid Electrical Vehicle: Powertrain Optimization Strategy and Experiment”, Ph.D. Thesis, University of Minnesota, 2017.
Copyright © 2017 by ASME
View article in PDF format.

References

United States Environmental Protection Agency, “Sources of Greenhouse Gas Emissions”, https://www.epa.gov/ghgemissions/sources-greenhouse-gas-emissions, accessed June 10, 2017.
United States Department of Transportation, “Average Fuel Efficiency of U.S. Light Duty Vehicles”, https://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_04_23.html, accessed June 10, 2017.
Hu, J., Shao, Y., Sun, Z., Wang, M., Bared, J. and Huang, P., “Integrated Optimal Eco-Driving On Rolling Terrain for Hybrid Electric Vehicle with Vehicle-Infrastructure Communication”, Transportation Research Part C, 68, 2016, pp. 228– 244. [CrossRef]
Mohd Zulkefli, M.A., “Connected Hybrid Electrical Vehicle: Powertrain Optimization Strategy and Experiment”, Ph.D. Thesis, University of Minnesota, 2017.

Figures

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In