关键词: 智能汽车；操纵稳定性；混合逻辑动态模型；HMPC

Abstract: To improve the accuracy and adaptability of lateral control for intelligent vehicles under different working conditions, a hybrid model predictive control (HMPC) strategy based on the mixed logical dynamical (MLD) model is proposed. The dynamics model is established using the piecewise affine approach, integrated with a two-degree-of-freedom vehicle dynamics model, and the hybrid logic dynamic model of the system is constructed via the hybrid systems description language (HYSDEL). On this basis, an HMPC controller is designed, which transforms the tracking control problem into a mixed integer quadratic programming problem for solution. Simulations are conducted on a co-simulation platform combining MATLAB/Simulink and CarSim under doublelane-change conditions. The results show that the designed hybrid model predictive controller can stably track the reference trajectory even when the system exhibits strong nonlinearity. Especially under low-adhesion road conditions, compared with the conventional model predictive control (MPC), the HMPC strategy reduces the root mean square error of the yaw rate from 1.56 rad·s -1 to 0.88 rad·s -1 , a decrease of about 43.6%, and reduces the root mean square error of the lateral speed from 1.43 m·s -1 to 0.65 m·s -1 , a decrease of about 54.5%. The controller maintains stable tracking under strongly nonlinear conditions, effectively suppresses yaw and lateral speed fluctuations, and enhances handling stability and tracking accuracy in complex working conditions.

Key words: intelligent vehicle; handling stability; mixed logical dynamical model; HMPC