Research On The Localization Method Of Supervision Robot Based On Multi-sensor Combination

Under the background of the great development of computer technology and artificial intelligence technology,intelligent robot,as the product of the interdisciplinary of computer and artificial intelligence,has made great progress,from laboratory to practical application.Among them,the supervision robot is a typical case of intelligent robot moving towards practical application.The supervision robot can replace the police to complete high-intensity and repetitive inspection work at night,and can avoid the safety accidents caused by police fatigue,illegal inspection and other factors.The robot completely replaces the police to complete patrol detection and other work,and also puts forward higher requirements for the robot itself.First of all,the environment in which the robot operates is the corridor of the supervision office,most of which are long and straight corridors with short positioning shape and long longitudinal length;secondly,the robot needs to stay at the door of each supervision office for a certain period of time for detection;finally,the robot needs to patrol continuously at night,that is,the robot needs to run continuously without restart.All of the above requirements require the robot to have a good positioning ability,and the robot also needs to have a good error elimination ability,otherwise the robot will grow with the running time and the accumulated error will continue to increase,which will not be able to complete the positioning task in the later stage.In order to solve this problem,this paper will study the precise positioning algorithm of the robot in the prison environment,in order to get a good scheme to effectively improve the positioning accuracy of the robot.This paper is mainly from the following aspects:1.Firstly,it introduces the hardware and software of the robot in the supervision institute.In the hardware aspect,it mainly introduces the structure of the robot chassis and the sensors needed for positioning.In the software aspect,it mainly introduces the overall framework of the software and the functions of each module.2.The classical Bayesian theory of robot positioning algorithm is analyzed,and the Kalman filter,extended Kalman filter and particle filter based on Bayesian filter algorithm are analyzed.3.Location algorithm.Based on the research of the positioning algorithm and the analysis of the positioning method of the robot,a high-precision positioning algorithm for the specific positioning needs of the robot is proposed.The algorithm is based on the combination of lidar,wheel odometer and monocular camera.Firstly,the coarse location is done based on lidar and wheel odometer.Monocular camera corrects the location and removes the accumulated error by identifying the artificial identification in the environment.4.Finally,a test platform is built for the proposed method to verify the effectiveness of the proposed positioning method.Experiments show that this method can ensure that the robot has a good positioning accuracy when it is in a single monitoring room.At the same time,when it is running for a long time,the robot has a good ability of error elimination,and the positioning accuracy can still meet the needs of accurate positioning.