A study on actuator delay compensation using predictive control technique with experimental verification

In this paper, a model predictive control (MPC) method for actuator delay compensation in vehicle stability control is presented. At each sampling time, an MPC controller computes required torque adjustments in order to follow desired vehicle dynamics on slippery road condition. Actuator delay in a control loop can diminish controller effectiveness to a great extent and may even engender instability in a critical driving situation. Two control structures with different computational complexities are presented. In the first approach, the MPC problem is formulated such that actuator dynamics is considered in prediction model. Whereas, the second approach investigates actuator lag impact on control action allocation (distribution) through a transfer matrix. In contrast to common delay handling approaches, the proposed method is simply real-time implementable and does not necessitate an intricate design procedure. The complication of model predictive controller remains intact, since online computation load is not substantially altered. Discussions are presented on computational burden and performance of two control schemes using computer-aided simulations in MATLAB/CarSim environment as well as experiments. A rear-wheel-drive sport utility vehicle equipped with differential braking is utilized for numerical simulations and experimentations. According to the simulation results, success of the proposed method in tackling actuator delay in vehicle stabilization is inquired. Experimental tests also demonstrate competency of the proposed technique. Although this technique is applied to a vehicle stability control with a particular actuator, it is not limited to a certain set of problems and can be employed for a broad variety of actuators or more generally, model predictive control problems with actuator identifiable time delays.