基于PSO-Newton的并联机构位置正解研究

The Research of Parallel Mechanism Forward Position based on PSO-Newton

并联机构位置正解可以转化为非线性方程组问题,针对这类问题的复杂性和多解性,提出将共享适应度粒子群算法和Newton迭代法相结合的求解非线性方程组的新方法——PSO-Newton算法。该方法使用共享适应度机制引导算法寻找全部最优值,以全局最优值为初始点进行Newton迭代,极大提高了算法的收敛速度,变异算子的引入保证个体多样性有助于提高全局寻优性能。最后将该算法应用于2PR&2SPS并联机器人机构位置正解问题,结果表明该算法效果良好,是一种快速求解并联机构位置正解的新算法。

The solution of the forward position of parallel mechanism can be transformed into the problem of nonlinear equations.Aiming at the complexity and multi-result of such problem,a new method for solving nonlinear equations —PSO-Newton algorithm combined sharing fitness of particle swarm optimization and Newton iteration is proposed.The fitness sharing mechanism is used to guide the algorithm finding all the global optimum value.Taking goal best value of PSO as the initial point for Newton iteration,the convergence speed of the algorithm is improved greatly.The diversity of individuals is ensured by introducing mutation operators,and contributing to improve the global optimization performance.At last,the algorithm is applied to the forward position of the 2PR2SPS parallel robotic mechanism position,and the result shows that the algorithm works well.It is a new algorithm for fast solving the parallel mechanism forward position.