A parameter-optimised rule-based control strategy for front-rear torque vectoring in electric vehicles with multiple motors and disconnect clutches

Applying a proper front/rear torque distribution in electric vehicles with multiple motors leads to battery energy consumption reduction, and thus the vehicle range extension. The energy efficiency can further be enhanced by equipping the e-motors with disconnect clutches, because the drag loss can be avoided by disconnecting and stopping the inactive motors to zero speed. In this paper, a practical parameter-optimized rule-based (RB) torque vectoring control strategy is proposed, which manipulates front/rear torque distribution ratio and clutch state control inputs to minimize the energy consumption. The strategy accounts for clutch connect transient losses and a requirement on limiting the dog clutch switching frequency for improved durability. It relies on a set of two-wheel drive/all-wheel drive switching curves with associated hysteresis and corresponding control rules. The RB performance is verified against a more sophisticated model predictive control (MPC) strategy and a globally optimal, dynamic programming (DP)-based offline control trajectory optimization benchmark. The DP optimization results reveal that the energy consumption reduction achieved through the disconnect clutch functionality is around 6%, on top of up to 5% reduction achieved by torque distribution itself. The RB strategy closely approaches the DP benchmark, it outperforms the MPC strategy fed by realistic, error-prone vehicle velocity predictions, and is robust with respect to driving cycle features.