Research on Path Tracking Method of Electrified Robot Based on Deep Neural Network

Abstract

Robot tracking is widely used in industrial live work, but most of the tracking algorithms are not effective in complex scene tracking. It is difficult to realize accurate recognition and tracking only by traditional target detection and tracking algorithm. This paper studies the mechanism of the influence of the electromagnetic interference on the robot control of the high voltage transmission line, and proposes a multi feature fusion target robot tracking algorithm based on the deep learning technology. The algorithm proposes an improved model of yolov3 and deep sort mechanism, which can identify and track the targets in complex scenes in real time and accurately. The experimental results show that the proposed algorithm can effectively and accurately track the target robot in complex scenes, and the experimental results are improved significantly.