改进粒子群算法在LQR半主动悬架的应用

针对车辆半主动悬架LQR控制中Q矩阵和R矩阵往往由经验取值的问题，提出一种基于改进粒子群算法的LQR控制方法。该算法采用随机惯性权重代替了传统粒子群算法的固定惯性权重，提高了求解精度和效率，得到了更加具有适应性的LQR控制矩阵系数。为验证此方法的有效性，基于天棚阻尼模型建立1/4车被动悬架模型和半主动悬架模型，利用线性二次最优控制建立LQR控制器，并利用优化算法得到新的控制矩阵。通过仿真对比被动悬架、LQR控制的LQR半主动悬架、改进粒子群算法优化后的优化LQR悬架的各项性能参数，发现优化LQR悬架在悬架动挠度没有受到影响的前提下，使车辆的垂向加速度和轮胎动载荷得到有效降低，提高了车辆的行驶平顺性和操纵安全性。

Application of Improved Particle Swarm Algorithm in Vehicle LQR Semi-active Suspension

Aiming at the problem that the Q matrix and R matrix in the LQR control of vehicle semi-active suspensions are often valued by personal experience, an LQR control method based on an improved particle swarm algorithm is proposed. The algorithm uses random inertia weights instead of the fixed inertia weights of the traditional particle swarm algorithm, improves the accuracy and efficiency of the solution, and obtains more adaptive LQR control matrix coefficients. In order to verify the effectiveness of this method, a quarter-car passive suspension model and a semi-active suspension model are established based on the ceiling damping model, the LQR controller is established using linear quadratic optimal control, and the new control matrix is obtained using the optimization algorithm. Through simulation and comparison of various performance parameters of passive suspension, LQR semi-active suspension controlled by LQR, and optimized LQR suspension optimized by improved particle swarm algorithm, it is found that the optimized LQR suspension effectively reduces the vertical acceleration of the vehicle and the dynamic load of the tire on the premise that the dynamic deflection of the suspension is not affected, and improves the driving comfort and handling safety of the vehicle.