Model Predictive Direct Motion Control for Distributed Drive Electric Vehicles

Abstract

For the motion of distributed drive electric vehicles, this paper deigns a novel motion control strategy based on model predictive control, which can control the longitudinal and lateral motion simultaneously. Different with traditional strategies containing one vehicle controller and four motor controllers, this strategy treats the vehicle and its in-wheel motors as one control object and uses only one controller to realize motion control, so that control costs are reduced and the handling performance is improved. First, an improved unified model of the vehicle and in-wheel motors is established to predict future states. Considering different time constants in vehicle dynamic system and motor electrical system, coupling predictions are analyzed in detail and a linear model is proposed to simplify the complex computation. Then, based on the fact that inverters of in-wheel-motor systems have discrete and finite voltage vectors, a model predictive direct motion control strategy is proposed. A cost function is designed to find out optimal voltage vectors which make vehicle follow the desired motion and meet constrains. Finally, to deal with huge amount of calculation and over-actuation problem caused by four in-wheel-motor systems, a deadbeat solution is applied to optimize the optimal voltage vectors calculation. Simulations under the proposed and traditional strategies are carried out, in which that the proposed strategy can realize the vehicle motion control well in different directions, and improve the vehicle handling performance. A practical implementation method is also provided experimentally to realize the strategy on four-motor system.

Keywords

• Distributed drive electric vehicle
• in-wheel motor
• model predictive control
• vehicle motion control