Research And Application Of Autonomous Localization Technology Of Inspection Robot Based On Laser Radar

With the development of technology,robots have appeared in more and more fields.A demand for robots is more urgent in industries with simple,repetitive and high-risk characteristics.For example,in substations and cable tunnels,it is necessary to monitor the operation of power equipment at all times.The traditional monitoring method is manual inspection,but which is low in efficiency,high in cost,slow in data updating and harmful to the inspectors.Therefore,it is a general trend to replace manual inspection with robot inspection.According to whether there is a track,inspection robots can be divided into rail type and trackless type.Former has poor flexibility and is difficult to cater to the rapidly growing performance requirements.Therefore,the trackless inspection robot has become a research hotspot.One of its technical difficulties is robot localization.At present,the localization of trackless inspection robot has some problems,such as low accuracy,strong dependence on the environment,and poor stability.As inspection robots designed towards more intelligent,workers can complete various inspection tasks only in an office environment,which greatly saves costs and improves efficiency.In order to enable the inspection robot to efficiently complete the inspection task,this thesis,from the perspective of the high accuracy and stability localization of the robot,carries out a research on the autonomous localization technology of the inspection robot based on the laser radar,and applies it to the practical engineering.The main contents of this thesis are as follows:(1)A localization model of the robot,an analysis of factors affecting localization accuracy and a localization system scheme are constructed.Firstly,based on characteristics of robot kinematics and localization sensor,a localization model of robot is constructed.Secondly,factors affecting the localization accuracy are analyzed with this model.Finally,a localization system scheme is designed for the inspection environment and the factors affecting the localization accuracy.(2)Localization technology of robot is analyzed and realized.In this thesis,for the characteristics of relative and absolute localization errors,UMBmark(University of Michigan Benchmark test)experimental algorithm is used to calibrate the odometer system error.In addition,a data fusion algorithm combines the data of the odometer and the Inertial Measurement Unit(IMU)to calibrate the odometer non-systematic error.Experimental results show that the relative localization accuracy is improved by the two error calibrations.Finally,based on the relative pose,adaptive Monte Carlo algorithm with improved proposed distribution is used to achieve absolute localization.Experimental results show that this algorithm makes the absolute pose more accurate.(3)An improved iterative closest point algorithm based on global map matching is proposed.Aiming at the accumulation of pose errors and pose drift of robots,an improved iterative closest point algorithm based on global map matching is proposed to correct pose errors.Experimental results show that the matching effect of this algorithm is ideal,and the localization accuracy of the robot is improved again.(4)The evaluation criteria of robot localization performance is constructed,and the test results are analyzed.In this thesis,a more comprehensive evaluation criterion is designed by investigating the evaluation method of localization performance.Based on this criterion,a test scheme is designed.Finally,by comparing and analyzing the results of localization test,it is shown that the autonomous localization of robots with high accuracy and stability is realized.