Non-Linear Model Predictive Anti-Jerk Cruise Control for Electric Vehicles with Slip-Based Constraints

Abstract

Electric vehicles (EVs) are usually fitted with Cruise Control (CC) systems, an Advanced Driver Assistance System (ADAS), which regulates the speed of the vehicle in response to an acceleration input. The rate of acceleration is usually regulated by the traction controller. However, most traction controllers are on-off controllers and are only activated when slip exceeds the desired limits resulting in deterioration in the performance of the cruise controller. EVs have a much faster torque response as compared to conventional vehicles, resulting in jerk arising as a result of wheel slip or flexibility in the half-shaft. In this research, we develop a non-linear model predictive low-jerk cruise controller for an electric vehicle, which reduces jerk occurring due to halfshaft flexibility and wheel slip concurrently. A high-fidelity longitudinal dynamics model has been developed for the test vehicle for our research, a Toyota Rav4EV. A powertrain model based on Pacejka relaxation length tire model has been used to study the slip response characteristics. The jerk performance of the controller has been assessed using the high-fidelity vehicle model while following a US06 driving cycle. The real-time capability of the MPC controller has been demonstrated through Hardware-in-the-loop (HIL) experiments.