Abstract: Based on hybrid particle swarm optimization, a linear optimal controller for vehicle active suspension is designed to reduce the Bodywork Acceleration（BA）, Suspension Dynamic Schedule（SWS） and Tire Dynamic Deflection（DTD）. Firstly, a 2-DOF dynamic model of a 1/4 vehicle active suspension is established. Then, the hybrid particle swarm algorithm is used to optimize suspension stiffness, suspension damping coefficient and weight matrix of LQG controller. Lastly, the model of different working condition is simulated and analysed under Matlab/Simulink environment. The simulation results illustrate that the riding comfort and handling stability of active suspension have been improved and the root mean square of BA, SWS and DTD is decreased by 22.56%, 44.27%, 19.75% after optimized by hybrid particle swarm.

Key words: hybrid particle swarm, Linear Quadratic Gaussian（LQG） controller, active suspension

摘要： 基于混合粒子群优化（Hybrid Particle Swarm Optimization，HPSO）算法设计了一种以降低车身加速度（BA），悬架动行程（SWS）和轮胎动位移（DTD）为目标的车辆主动悬架线性最优控制器。建立了2自由度1/4车辆主动悬架动力学模型，运用混合粒子群优化算法对LQG控制器的权值矩阵进行优化求解，在Matlab/Simulink环境下，对不同工况下的车辆悬架进行了仿真分析。仿真结果表明，经过混合粒子群算法优化后的主动悬架在行驶平顺性和操纵稳定性上有所改善，并且优化后主动悬架性能指标BA，SWS和DTD的均方根值最大分别减少了22.56%，44.27%和19.75%。

关键词: 混合粒子群算法, 线性二次型（LQG）控制器, 主动悬架