基于约束满足的多目标双足机器人动态步态参数优化*

目前的步态优化算法仅仅实现了对单一目标的优化，把双足机器人步态优化看做是多目标优化问题，构建了衡量稳定性、能量消耗、步行速度三个目标评价函数。考虑到直接对多个目标加权求和的方法不能很好地处理多目标问题,提出一种新的基于约束满足的多目标步态参数优化算法，其思想是把基于惩罚函数的SPEA2(strength Pareto evolutionary algorithm2 )应用到多目标双足机器人动态步态参数优化问题上，规划出了同时满足这三个目标的动态优化步态。通过仿真实验表明了算法的有效性。

Multi-objective dynamic gait parameters optimization for biped robot based on constraint satisfaction

Current gait optimization algorithm only achieved a single objective optimization, this article took the biped robot gait optimization as multi-objective optimization problem,constructed three objective evaluation functions about stability,power consumption and walking speed. Considering the weighted summation of fitness functions can not handle commendably the problem about multiple conflicting objectives, this paper proposed a new kind of multi-objective gait parameter optimization algorithm based on constraint satisfaction, Its idea is that applying SPEA2 based on penalty function to multi-objective biped robot dynamic gait parameter optimization problem. And planning out dynamic optimization gait meeting the three conflicting objectives. The simulation experiments show the validity of this algorithm.