Advanced Multi-Mode Control Strategy for a Parallel Hybrid Electric Vehicle Based on Driving Pattern Recognition

Abstract

The adaptive multi-mode control strategy (AMMCS) is defined as the control strategy that switches control parameters for the purpose of adjusting vehicles to diverse traffic conditions and driver’s habits. This strategy is composed of off-line and on-line procedures. In the off-line procedure, several sets of control parameters are optimized under representative driving patterns (RDP). In the on-line procedure, the control parameter switching or interpolation is periodically activated based on the driving pattern recognition (DPR) algorithm, assuming that the driving pattern during the future control horizon doesn’t change significantly compared to the past pattern. The AMMCS is conceptually similar to one of predictive control theories, namely the receding horizon control which is also known as model predictive control. The AMMCS is expected to be applied well to hybrid electric vehicle (HEV) system which is very sensitive to driving patterns. Furthermore, the AMMCS can be combined with the two conventional control strategies using global and local optimization techniques to improve performances further. The design goal of the AMMCS is to minimize fuel consumption and NOx for a pre-transmission single shaft parallel HEV.

Route-Based Control of Hybrid Electric Vehicles

“…Another approach subcategory relies on repeated cycle learning to tune control parameters. Literature examples of the first approach subcategory show modest improvements with route-based control that are limited in part by the fact that past operation does not always correctly indicate the upcoming driving type [10,11]. The second approach subcategory suggests greater fuel efficiency gains but entails increased cycle sensitivity by requiring several sequential repetitions of the same driving pattern to achieve control parameter learning [12].…”