Abstract: A simplified model of three-link underactuated planar robot is created, and the dynamic equation and integral feature of the model is analyzed. The method of model degradation is applied to calculate the relationship of angle between the joints constraint by reducing the partly integrable equation of the model to two completely integrable equations. To overcome some defects of particle swarm algorithm of which slow convergence rate for inverse kinematics and easily falling into a local optimum, particle swarm optimization neural network learning algorithm is proposed, which is based on the joint angle constraints and a fitness function which is the sum of the squares of the errors between practice and ideal angle. The method is proved to be effective in the simulation experiments.

Key words: underactuated planar robot, inverse kinematics, particle swarm, neural network, angle error

摘要： 针对欠驱动平面机器人三连杆简化模型，分析模型的动力学方程及其积分特性，采用模型退化方法，将该模型的方程式从部分可积降阶为两个完全可积的子方程式，计算关节之间的角度约束关系。为克服粒子群算法实现逆运动学求解收敛慢、易陷入局部最优等缺陷，依据得到的关节角度约束关系，以实际角度与理想角度的误差平方和作为适应度函数，提出利用粒子群优化神经网络的学习型算法进行求解。仿真实验验证了方法的有效性。

关键词: 欠驱动平面机器人, 逆运动学, 粒子群, 神经网络, 角度误差