The Key Technologies Research Of SLAM For Millimeter-wave Radar

With the continuous rise and development of artificial intelligence technology,whether in UAV or unmanned driving and other related fields,it is particularly important for the subject of motion to observe environmental information and determine its own position through sensors in an unknown environment.Because traditional location and mapping relies extensively on vision and lidar,it lacks adaptability to harsh environments such as darkness,smoke and haze.However,millimeter-wave radar has a good application prospect because of its high measurement accuracy and not easy to be interfered by the environment.Therefore,how to use millimeter-wave radar to process and optimize its data,and then apply it to location and mapping has been a potential research topic in recent years.After investigating and analyzing the different conditions and requirements for robot localization and mapping at home and abroad,the thesis combinates with the millimeter wave radar sensor and data processing technology,designs a millimeter wave radar positioning system which can filter and process data,and suitable for different environment.At the same time,the system used robot operating system as the basic development platform,can be used in smoke,dust,darkness and others harsh environment.Here,the main research contents of this thesis are as follows:(1)The research status of robot positioning and mapping technology at home and abroad are investigated by this thesis.After analyzing the development trend of millimeter-wave radar development and application,the research content is determined.(2)In this thesis,the parameters of several radar sensors are analyzed and compared.IWR1443 of Texas Instruments is selected as the main chip of the millimeter wave radar sensor for the requirements of positioning and mapping.According to the data interaction principle the millimeter wave radar and the robot car,the overall system scheme is determined.(3)In this thesis,the receiving,processing and optimization algorithms of millimeter-wave radar signal are studied.According to the disadvantages of millimeterwave radar receiving point cloud signal,such as small data volume,small measurement range and easy to be interfered by reflection,the rotating steering gear and a variety of filtering processing technologies are adopted to optimize the signal.At last a better acquisition effect is obtained.(4)The basic principle of data interaction between the robot car and the upper computer is studied,and the detection effect is improved by multi-sensor fusion.Then,based on the millimeter-wave radar signals in the robot operating system used for positioning and mapping algorithms are studied.The millimeter-wave radar point cloud data collected are constructed using mapping algorithms based on octree and occupied grid to achieve positioning and mapping for different environments.Finally,experiments are carried out to verify the results.In summary,this thesis studies the key technologies of positioning mapping based on millimeter wave radar.First,the 76?81GHz multi-input multi-output millimeter-wave radar is selected according to the demand,and the accuracy of the point cloud is ensured by the combination of through,radius and statistical filtering algorithms.By rotating scanning,the radar viewing angle and the number of point clouds are expanded,and the data fusion with the visual sensor increases the accuracy of obstacle judgment.Finally,the robot car combined with the robot operating system realized the verification of the millimeter wave radar localization and mapping technology in the simulated harsh environment experiment.