【摘要】 为解决机器人被动行走研究中不动点搜索的时间和结果都依赖于初始值的问题,该文研究了形成周期性步态的充要条件,得到不动点处状态变量之间的约束关系,从而将搜索空间从二维降低为一维,并提出了一维空间中不动点搜索的新算法。该算法能在固定的搜索时间内一次性找到所有稳定和不稳定的周期性步态的不动点,而不需要进行初始值的选择。该文通过数值仿真验证了新算法的有效性。更多还原

【Abstract】 Both the search time and the search result for fixed points during passive walking depend on the initial values.This paper investigates the necessary and sufficient conditions for periodic gaits to obtain constrains between the state variables at the fixed points,and reduce the two dimensional search space to a one dimensional space with an algorithm,which was then used to search for fixed points in the one dimensional pace.This algorithm finds all the stable and unstable fixed points during periodic gaits by one search in a fixed time period without specification of initial values.The algorithm is validated by simulations.更多还原