0
Research Papers

Real-Time Implementation and Hardware Testing of a Hybrid Vehicle Energy Management Controller Based on Stochastic Dynamic Programming

[+] Author and Article Information
Daniel F. Opila

Graduate Research Assistant
Department of Mechanical Engineering,
University of Michigan,
Ann Arbor, MI 48109
e-mail: dopila@umich.edu

Xiaoyong Wang

e-mail: xwang67@ford.com

Ryan McGee

e-mail: rmcgee3@ford.com
Ford Motor Company
Research and Advanced Engineering,
Dearborn, MI 48120

J. W. Grizzle

Professor
Department of Electrical Engineering,
University of Michigan,
Ann Arbor, MI 48109
e-mail: grizzle@umich.edu

Contributed by the Dynamic Systems Division of ASME for publication in the Journal of Dynamic Systems, Measurement, and Control. Manuscript received November 2, 2010; final manuscript received May 28, 2012; published online November 7, 2012. Assoc. Editor: Marcelo J. Dapino.

J. Dyn. Sys., Meas., Control 135(2), 021002 (Nov 07, 2012) (11 pages) Paper No: DS-10-1320; doi: 10.1115/1.4007238 History: Received November 02, 2010; Revised May 28, 2012

An energy management controller based on shortest path stochastic dynamic programming (SP-SDP) is implemented and tested in a prototype vehicle. The controller simultaneously optimizes fuel economy and powertrain activity, namely gear shifts and engine on–off events. Previous work reported on the controller's design and its extensive simulation-based evaluation. This paper focuses on implementation of the controller algorithm in hardware. Practical issues concerning real-time computability, driver perception, and command timing are highlighted and addressed. The SP-SDP controllers are shown to run in real-time, gracefully handle variations in engine start and gear-shift-completion times, and operate in a manner that is transparent to the driver. A hardware problem with the test vehicle restricted its maximum engine torque, which prevented a reliable fuel economy assessment of the SP-SDP controller. The data that were collected indicated that SP-SDP controllers could be straightforwardly designed to operate at different points of the fuel economy tradeoff curve and that their fuel economy may equal or exceed that of a baseline industrial controller designed for the vehicle.

Copyright © 2013 by ASME
Your Session has timed out. Please sign back in to continue.

References

Sciarretta, A., and Guzzella, L., 2007, “Control of Hybrid Electric Vehicles,” IEEE Control Syst. Mag., 27(2), pp. 60–70. [CrossRef]
Paganelli, G., Tateno, M., Brahma, A., Rizzoni, G., and Guezennec, Y., 2001, “Control Development for a Hybrid-Electric Sport-Utility Vehicle: Strategy, Implementation and Field Test Results,” Proceedings of American Control Conference (ACC), pp. 5064–5069. [CrossRef]
Paganelli, G., Ercole, G., Brahma, A., Guezennec, Y., and Rizzoni, G., 2001, “General Supervisory Control Policy for the Energy Optimization of Charge-Sustaining Hybrid Electric Vehicles,” JSAE Rev., 22(4), pp. 511–518. [CrossRef]
Lin, C.-C., Peng, H., Grizzle, J., Liu, J., and Busdiecker, M., 2003, “Control System Development for an Advanced-Technology Medium-Duty Hybrid Electric Truck,” Proceedings of the International Truck and Bus Meeting and Exhibition, Fort Worth, TX.
Boyali, A., Demirci, M., Acarman, T., Guvenc, L., Tur, O., Ucarol, H., Kiray, B., and Ozatay, E., 2006, “Modeling and Control of a Four Wheel Drive Parallel Hybrid Electric Vehicle,” Proceedings of IEEE International Conference on Control Applications, pp. 155–162. [CrossRef]
Graham, L., Christenson, M., and Karman, D., 2006, “Light Duty Hybrid Vehicles—Influence of Driving Cycle and Operating Temperature on Fuel Economy and GHG Emissions,” Proceedings of the IEEE EIC Climate Change Technology, M.Christenson, ed., pp. 1–6. [CrossRef]
Kermani, S., Delprat, S., Trigui, R., and Guerra, T. M., 2008, “Predictive Energy Management of Hybrid Vehicle,” Proceeding of IEEE Vehicle Power and Propulsion Conference, pp. 1–6. [CrossRef]
Kermani, S., Delprat, S., Guerra, T. M., and Trigui, R., 2009, “Predictive Control for HEV Energy Management: Experimental Results,” Proceedings of IEEE Vehicle Power and Propulsion Conference, pp. 364–369. [CrossRef]
Tate, E., Grizzle, J., and Peng, H., 2008, “Shortest Path Stochastic Control for Hybrid Electric Vehicles,” Int. J. Robust and Nonlinear Control, 18, pp. 1409–1429. [CrossRef]
Bertsekas, D. P., and Tsitsiklis, J. N., 1996, Neuro-Dynamic Programming, Athena Scientific, Belmont, MA.
Bertsekas, D., 2005, Dynamic Programming and Optimal Control, Vol. 1, Athena Scientific, Belmont, MA.
Bertsekas, D., 2005, Dynamic Programming and Optimal Control, Vol. 2, Athena Scientific, Belmont, MA.
Johannesson, L., Asbogard, M., and Egardt, B., 2007, “Assessing the Potential of Predictive Control for Hybrid Vehicle Powertrains Using Stochastic Dynamic Programming,” IEEE Trans. Intell. Transp. Syst., 8(1), pp. 71–83. [CrossRef]
Lin, C.-C., Peng, H., and Grizzle, J., 2004, “A Stochastic Control Strategy for Hybrid Electric Vehicles,” Proceedings of the American Control Conference, pp. 4710–4715.
Lin, C.-C., Peng, H., Grizzle, J., and Kang, J.-M., 2003, “Power Management Strategy for a Parallel Hybrid Electric Truck,” IEEE Trans. Control Syst. Technol., 11(6), pp. 839–849. [CrossRef]
Opila, D., 2010, “Incorporating Drivability Metrics Into Optimal Energy Management Strategies for Hybrid Vehicles,” Ph.D. thesis, University of Michigan, Ann Arbor, MI.
Opila, D., Wang, X., McGee, R., Gillespie, R., Cook, J., and Grizzle, J., 2012, “An Energy Management Controller to Optimally Trade Off Fuel Economy and Drivability for Hybrid Vehicles,” IEEE Trans. Control Syst. Technol., 20(6), pp. 1490–1505. [CrossRef]
Belton, C., Bennett, P., Burchill, P., Copp, D., Darnton, N., Butts, K., Che, J., Hieb, B., Jennings, M., and Mortimer, T., 2003, “A Vehicle Model Architecture for Vehicle System Control Design,” Proceedings of SAE World Congress and Exhibition, Paper No. 2003-01-0092. [CrossRef]
McGee, R. A., 2003, “Model Based Control System Design and Verification for a Hybrid Electric Vehicle,” SAE Technical Paper Series, Future Transportation Technology Conference, Costa Mesa, CA, June 2003, SAE Paper No. 2003-01-2308. [CrossRef]
Ramaswamy, D., McGee, R., Sivashankar, S., Deshpande, A., Allen, J., Rzemien, K., and Stuart, W., 2004, “A Case Study in Hardware-in-the-Loop Testing: Development of an ECU for a Hybrid Electric Vehicle,” 2004 World Congress and Exhibition, Detroit, MI, SAE Paper No. 2004-01-0303. [CrossRef]
Tate, E. D., Jr., 2007, “Techniques for Hybrid Electric Vehicle Controller Synthesis,” Ph.D. thesis, University of Michigan, Ann Arbor, MI.
Opila, D., Aswani, D., McGee, R., Cook, J., and Grizzle, J., 2008, “Incorporating Drivability Metrics Into Optimal Energy Management Strategies for Hybrid Vehicles,” Proceedings of IEEE Conference on Decision and Control, pp. 4382–4389. [CrossRef]
Opila, D., Wang, X., McGee, R., Cook, J., and Grizzle, J., 2009, “Performance Comparison of Hybrid Vehicle Energy Management Controllers on Real-World Drive Cycle Data,” Proceedings of American Control Conference (ACC '09), pp. 4618–4625. [CrossRef]
Opila, D., Wang, X., McGee, R., Cook, J., and Grizzle, J., 2009, “Fundamental Structural Limitations of an Industrial Energy Management Controller Architecture for Hybrid Vehicles,” Proceedings of ASME Dynamic Systems and Control Conference, pp. 213–221. [CrossRef]

Figures

Grahic Jump Location
Fig. 1

The prototype hybrid: a modified Volvo S-80

Grahic Jump Location
Fig. 2

Fuel economy and engine activity for simulation and hardware testing on the Federal Test Procedure (FTP72). A component failure limited engine torque for all hardware testing, resulting in decreased fuel economy. All results are normalized to the simulated baseline controller.

Grahic Jump Location
Fig. 3

Vehicle configuration

Grahic Jump Location
Fig. 4

The overall development process

Grahic Jump Location
Fig. 5

The increasing complexity of controller testing in this work

Grahic Jump Location
Fig. 6

Multirate actuator commands

Grahic Jump Location
Fig. 7

One of the early HIL simulations on the New European Drive Cycle. The vehicle speed, gear, and engine state show reasonable behavior. The automated driver model was not well-tuned at this point, so the velocity tracking shows some lag and overshoot.

Grahic Jump Location
Fig. 8

Illustration of the command update scheme. Controls are organized to easily permit updates at multiple rates while respecting appropriate constraints. The hashed section represents the selected gear at each update.

Grahic Jump Location
Fig. 9

The first driving attempt in the vehicle, which corresponds to the first three hills of NEDC. The difference between targeted and actual speed is due to driver inexperience. Accurately following speed traces on a chassis roll dynamometer is an acquired skill.

Grahic Jump Location
Fig. 10

Engine torque and driver pedal commands. The upper plot shows the original SP-SDP engine torque command as a dotted line (red) oscillating at relatively constant pedal. The solid (blue) line shows the command after this problem was fixed with a “bowl” penalty. Both commands are the raw output of the SP-SDP algorithm before low-pass filtering. The bottom plot shows accelerator pedal command in percentage of full range.

Grahic Jump Location
Fig. 11

The value function during an unexpected pedal oscillation. The recorded vehicle data are shown on the left for controller updates at six consecutive time steps. The total expected cost for each possible engine torque command is shown as a heavy solid line (red). The minimizing torque selected by the controller is indicated by the vertical dashed line (black) and demonstrates the cause of the oscillations shown in Fig. 10. This oscillation is removed by adding a “bowl” penalty on engine torque which adds a cost for torque changes. The column on the right represents this improved control response applying Eq. (6) to the same vehicle data. The bowl penalty is marked with a solid line and circles (blue) at the bottom of each plot. The bowl penalty is centered at the last commanded torque and visibly changes position from t = 16.8 s to t = 17.2 s due to the change in torque command. The minimizing torque selection no longer oscillates.

Grahic Jump Location
Fig. 12

Additional penalty added to the value function based on the change in engine torque. This is termed a “bowl” penalty due to its shape.

Grahic Jump Location
Fig. 13

Driving the FTP72 cycle with the baseline controller and two different SP-SDP controllers

Grahic Jump Location
Fig. 14

Driving the NEDC with the baseline controller and two different SP-SDP controllers

Grahic Jump Location
Fig. 15

Engine torque-speed operating points demonstrating the effects of a hardware failure. The plots on the left (15(a) and 15(c)) show the baseline controller, while the plots on the right (15(b) and 15(d)) show the SP-SDP controller. The top plots show the commanded torques, while the bottom plots show the delivered torque. The engine control computer clips the delivered torque at about 170 Nm.

Grahic Jump Location
Fig. 16

Detailed actuation traces for one “hill” of NEDC

Grahic Jump Location
Fig. 17

Zoom of the engine start dynamics. Note the 1.5 s delay between the parallel mode request and actual clutch engagement. EM1 is attached to the crankshaft and provides engine starting torque.

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