A real-time hybrid model predictive coordinated control strategy for extreme path tracking of 4WIDAVs

Vehicle maneuvers in extreme conditions involve highly nonlinear tire dynamics and strongly coupled multi-domain behaviors, making accurate modeling and fast, reliable control particularly challenging. To this end, this paper proposes a hybrid model predictive coordinated control (hMPCC) strategy that balances the trade-off between internal modeling accuracy and the fast implementation of the optimal control framework. The vehicle nonlinear dynamics are firstly approximated using a three-dimensional piecewise affine (PWA) system, forming the predictive-based hybrid system model. Then the relationship between the PWA system and tire conditions (linear, nonlinear, and saturation) is further assessed along with phase portrait analysis establishing the control mode reference for vehicle agility and stability. Based on the control mode reference, torque vectoring is optimised by synthesising an augmented adaptive weight tuning strategy (A-AWTS) with the proposed hybrid model predictive control. Extensive simulations in CarSim on the Nürburgring Nordschleife high-speed track demonstrate the algorithm’s superior control performance and high computational efficiency. The hMPCC achieves a smaller lateral offset and a root mean square heading error within 10% of that of high-precision nonlinear MPC, along with perfect longitudinal speed tracking performance. Furthermore, over ten closed-loop simulations, the hMPCC maintains average and maximum computational times within 3% and 1% of those for the time-efficient decoupled control.