论文部分内容阅读
仿人机器人的全方位步行参数对其行走稳定性、灵活性、快速性具有较大影响,然而物理机器人与描述其前后步幅连接约束的简化动力学模型间的数学关系却难于建立,因而难于获得优化目标表达式和相应的优化方法.本文从步幅跟随规划算法中提取出7个关键影响参数,并将标准实验工况下的步幅跟随性能指定为优化目标,从而将问题转化为一个黑盒优化过程.基于动力学仿真建立Kriging代理模型,通过Latin超立方初始实验和EGO(effective global optimization)迭代建模优化求解该问题.动力学仿真结果表明,在较少的实验代价下,该方法实现了全方位步行参数的优化,该方法能够实现步行速度和步幅跟随能力的综合提升.
The omni-directional walking parameters of a humanoid robot have a great influence on the walking stability, flexibility and rapidity. However, the mathematical relationship between a physical robot and a simplified dynamic model describing the connection constraint between its front and back steps is difficult to set up, Get the optimization objective expression and the corresponding optimization method.This paper extracts seven key influence parameters from the stride following programming algorithm and designates the stride following performance under the standard experimental conditions as the optimization goal so as to transform the problem into one The black box optimization process is based on dynamic simulation to establish the Kriging proxy model and to solve the problem through Latin hypercube initial experiments and EGO (effective global optimization) iterative modeling.The dynamics simulation results show that with less experimental cost, The method realizes the optimization of all-direction walking parameters, and the method can realize the comprehensive improvement of walking speed and stride following ability.