Research On Control System Of Intelligent Anti Epidemic Robot

Since 2020,the COVID-19 epidemic has spread all over the world.In order to cope with the epidemic,reduce contact,and reduce the risk of cross-infection,the paper designed an anti epidemic robot with autonomous disinfection,fixed-point distribution functions and social distance detection.The intelligent anti epidemic robot can replace the anti-epidemic personnel to carry out large-scale autonomous disinfection in high-risk areas,such as hospitals and airports;it has the function of fixed-point distribution,and can realize unmanned distribution of medical materials and luggage;it can detect the distance between two people in real time.and urge the crowds to maintain at least a specific social distance,and warnings will be issued if the distance is less than that.The robot consists of GD32 main control board,embedded industrial computer,16-line laser radar and kinect camera.Through the GD32F405RGT6 chip,it controls the PWM output of the four motors,collects the data of the encoder and gyroscope,and interacts with the embedded industrial computer through the serial port.The ROS master fuses the collected imu data,then uses the Cartographer algorithm for 3D mapping,uses scan-map and loopback detection and positioning,and uses the ground plane fitting method to extract ground information from the 3D map.The combined ground information and obstacle information are simplified to generate a passable map.When performing the disinfection task,the robot uses the algorithm of full coverage path,plans the trajectory,and starts patrol disinfection;when performing the distribution task,the robot uses the planner path planner according to the set target point,calculates the distribution route,and plans the trajectory and speed.The 3D target detection method of PointRCNN is improved,and the feature extraction of the original point cloud in the first stage of the model is changed to the VoxelRPN feature extraction network,which improves the operation speed.The intelligent epidemic prevention robot detects human targets by collecting images and 3D point cloud information in real time,thereby judging the distance between people and promoting a reasonable social distance between people.According to the above design,the functions of auto disinfection and fixed-point distribution are realized;The pointrcnn model is revised and the inference time is shortened from 190 ms to 158ms;In a variety of environments,the proportion of disinfection coverage was above the 90%.