Design And Path Planning Research Of Patrol Fire Fighting Robot

With the development and progress of the national economy and science and technology,automation technology is gradually replacing manual labor.In some environments where dangerous goods,flammable and explosive materials are stored,the use of intelligent automation technology instead of manual inspection,timely warning and preliminary treatment when encountering fire,can effectively reduce labor costs and reduce casualties.This paper designs a patrol fire robot for indoor environment from the aspects of hardware structure and software system,which can perform autonomous mapping and navigation,automatic fire positioning and autonomous fire extinguishing.And the RRT* algorithm is improved,the path of the robot in the process of inspection and fire extinguishing is optimized,the time cost is saved,and experiments are carried out in real scenarios to prove its feasibility.First of all,for a specific indoor place,a flexible and portable inspection fire robot is designed,and the robot is built according to its simulation model,and the upper computer,lower computer,lidar,depth camera and other hardware equipment of the robot are selected,so as to realize the design and physical construction of the four parts of the robot’s drive,execution,sensing and control.Secondly,a fire detection system based on Open CV is designed,which can realize the automatic detection and positioning of fire by robots.The system first uses the flame detection algorithm based on color criterion to realize flame detection,and then uses Py QT5 to design and develop the fire detection system interface.Aiming at the problem that dynamic fire,static flame and other interference substances cannot be distinguished,a dynamic flame detection algorithm based on MOG is proposed,and finally the improved algorithm is implanted into the upper computer for experimental verification,which proves the reliability and accuracy of the system and realizes the fire detection function of the robot.Then,based on the limitations of the RRT* algorithm,an improved RRT* algorithm is proposed,which converges the sampling points in the elliptic subset and expands a random tree from the start and end points at the same time to realize bidirectional search.This algorithm improves the efficiency of the RRT* algorithm and solves the problem that the robot has a long search time and the path is not optimal during the inspection process.Finally,the robot simulation model is integrated into Rviz and Gazebo,and a simulation environment is built in Gazebo to realize the basic motion control of the robot through keyboard control.Design drawing nodes,positioning nodes,path planning nodes and navigation nodes to realize the mapping navigation and daily inspection functions of inspection fire robots.The improved RRT* algorithm proposed above is jointly experimented in the built simulation environment and real environment,which verifies the practicability of the algorithm,realizes the autonomous mapping and navigation of the robot,and proves its feasibility in the real environment.