A search and rescue robot device realization based on ROS operating system combined with visual algorithm

Abstract

Aiming at the problem of low accuracy of intelligent devices to find missing people, a search and rescue robot device based on ROS intelligent navigation system was proposed. SolidWorks was used to improve the new mechanical structure that can switch autonomously to different terrain and combined with Yolov5 deep learning algorithm to complete face recognition. Using the stm32 single-chip microcomputer, we adjusted the cascade PID parameters of the motor and simulation to realize the omnidirectional movement control of the robot. Through the analysis and verification of the experimental data, the adaptive ability of the improved robot platform to the terrain was significantly improved. The method can accurately complete the task of face recognition and autonomous navigation, which is practical and improves the accuracy of finding the missing people.