Abstract: In order to solve the stability problem of the quadruped robot when its center of mass deviates from the geometric center of the torso, this paper proposes an optimization method based on an improved particle swarm algorithm. Firstly, it uses an oscillator based on the Hopf model to build a central pattern generator（central pattern generator, CPG） network topology, through the trajectory planning of the foot to determine the relevant parameters of the CPG model, and optimize the coupling coefficient matrix between the CPG units to enable it to output the correct gait signal. Then it uses the adaptive weight adjustment particle swarm. The algorithm quickly finds the optimal parameter combination of the output model through continuous iteration, and solves the stability problem caused by the shift of the center of gravity. Finally, using Webots and MATLAB to conduct simulation experiments on the proposed optimization method, the simulation results prove that the method can quickly and effectively improve the motion stability of the quadruped robot under the condition of shifting the center of gravity.

Key words: quadruped robot, Hopf model, central pattern generator（CPG）, particle swarm optimization（PSO）, adaptive weight

摘要： 为解决四足机器人在其质心偏离躯干几何中心时的稳定性问题，提出了一种基于改进粒子群算法的优化方法。使用基于Hopf模型的振荡器搭建中枢模式发生器（central pattern generator，CPG）网络拓扑结构，通过对足端进行轨迹规划进而确定CPG模型相关参数，并对CPG单元间的耦合系数矩阵进行优化，使其能够输出正确的步态信号；之后采用自适应调整权重粒子群算法，通过不断迭代快速寻找输出模型的最优参数组合，解决由于重心偏移带来的稳定性问题。利用Webots和MATLAB对所提出的优化方法进行仿真实验，仿真结果证明该方法能够快速、有效地提高四足机器人在重心偏移情况下的运动稳定性。

关键词: 四足机器人, Hopf模型, 中枢模式发生器（CPG）, 粒子群算法（PSO）, 自适应权重