AUTHOR=Mei Mingming , Cheng Shuo , Mu Hongyuan , Pei Yuxuan , Li Bo TITLE=Switchable MPC-based multi-objective regenerative brake control via flow regulation for electric vehicles JOURNAL=Frontiers in Robotics and AI VOLUME=10 YEAR=2023 URL=https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2023.1078253 DOI=10.3389/frobt.2023.1078253 ISSN=2296-9144 ABSTRACT=

Recent investigations of the electric braking booster (E-Booster) focus on its potential to enhance brake energy regeneration. A vehicle’s hydraulic system is composed of the E-Booster and electric stability control to control the master cylinder and wheel cylinders. This paper aims to address the independent closed-loop control of the position and pressure as well as the maintenance of the pedal feel. To track both the reference signals related to piston displacement and the wheel cylinder pressure, an explicit model predictive control (MPC) is developed. First, the new flow model is introduced as the foundation for controller design and simulation. Next, in accordance with the operational conditions, the entire system is divided into three switchable subsystems. The three distributed MPCs are constructed based on the linearized subsystems, and a state machine is used to perform the state jump across the controllers. A linear piecewise affine control law can then be obtained by solving the quadratic program (QP) of explicit MPC. Afterwards, the non-linear extended Kalman filter including the recorded time-variant process noise is used to estimate all the state variables. The effectiveness of the explicit MPC is evidenced by the simulations compared with a single MPC in regenerative and dead-zone conditions. The proposed controller decreases the latency significantly by 85 milliseconds, which also helps to improve accuracy by 22.6%. Furthermore, the pedal feel remains consistent, even when factoring in the number of vibrations caused by the inherent hydraulic characteristic of pressure versus volume.