Action Recognition Method for Multi-joint Industrial Robots Based on End-arm Vibration and BP Neural Network

Abstract

Recognizing the motion of the multi-joint industrial robot from the measurement signal is helpful to link the test signal with the motion joint and improve the accuracy of state evaluation. A motion recognition method for multi-joint industrial robots based on end-arm vibration and Back Propagation (BP) neural network is proposed in this paper. A three-axis vibration sensor is installed on the last joint of the multi-joint industrial robot to obtain the vibration signals and then segment the acquired signal according to the length of time and extract the features, establish a feature matrix, train the network model through a single joint motion feature matrix, and finally identify the action corresponding to each small segment of the signal in the multi-joint motion of the robot through the model. The experimental results show that the proposed motion recognition method based on end-arm vibration and BP neural network has high practical value in action state recognition of multi-joint industrial robots.