基于MEA-BP神经网络的6DOF工业机器人逆运动学研究

针对6DOF工业机器人逆运动学求解存在计算量大、通用性差、有奇异性等问题,提出一种基于思维进化算法(MEA)优化BP神经网络的工业机器人逆运动学求解方法。在机器人工作范围内,随机生成若干组关节角度值,通过正运动学方程获得机器人末端连杆位置和姿态,以末端连杆位置和姿态作为模型输入,关节角度作为模型输出,通过对样本数据的训练确定模型参数。使用该模型进行机器人逆运动学求解,并与传统的基于BP和RBF神经网络的求解方法进行比较,结果表明:该求解方法精度高、泛化能力强。

Inverse Kinematics Study of 6DOF Industrial Robot Based on MEA-BP Neural Network

The inverse kinematics solution for 6DOF industrial robots has the problems of large computational complexity, poor versatility and singularity. An inverse kinematics solution for industrial robots was proposed based on mind evolutionary algorithm (MEA) optimization BP neural network to address these problems. Several sets of joint angle values were randomly generated within the working range of the robot. Then, the position of the robot end link was obtained by the forward kinematics equation, the end link position was used as model inputs, and the joint angle was taken as the output of model. The model parameters were determined by the training sample data, and the model was used to solve the inverse kinematics of the robot. Compared with the traditional BP and RBF neural network solving methods, the results show that the method has higher precision and strong generalization ability.