汽车转向非线性平衡点遗传算法求解及其改进

Genetic Algorithm for Solving Nonlinear Equilibrium Points of Automobile Steering and Its Improvement

针对汽车转向非线性平衡点求解问题,研究了遗传算法求解效果并提出改进方法.建立汽车转向二自由度模型,说明汽车转向非线性平衡点只能数值迭代求解的原因,构造适于智能优化方法的适应度函数,提出了确定可行求解范围的方法.在车速70 km/h、路面附着系数0.5和前轮转角0~0.2 rad的行驶条件下,应用遗传算法求解得到3个平衡点.通过比较大小两个转角的适应值曲面,说明遗传算法求解小转角平衡点不满足精度的原因,提出了遗传算法与BFGS(broyden-fletcher-goldfarb-shanno)拟牛顿法融合的求解流程.结果表明:融合求解流程可以求解满足精度要求的小转角平衡点,求解效率高于遗传算法,弥补了遗传算法单独求解的不足.

Aiming at the problem of solving nonlinear equilibrium points of automobile steering,the solving effect of the genetic algorithm was studied and an improved method was proposed.A two-degree-of-freedom model of automobile steering was established,and the reason why the nonlinear equilibrium points of automobile steering can only be solved by numerical iteration was explained,a fitness function suitable for the intelligent optimization method was constructed,and a method to determine the feasible solution range was proposed.Under the driving conditions of a vehicle speed of 70 km/h,road adhesion coefficient of 0.5 and a front wheel angle of 0~0.2 rad,three equilibrium points were solved by using genetic algorithm.By comparing the fitness value surfaces of large and small steering angles,the reason why the genetic algorithm can not meet the solving accuracy of small steering angle equilibrium point was explained,and the solution flow of the fusing genetic algorithm and BFGS quasi-Newton method was presented.The results show that the proposed fusion solution process can solve the small steering angle equilibrium points that meet the solving accuracy requirements and have higher solving efficiency,which makes up for the shortage of the genetic algorithm solving alone.