Research On SLAM Technology In Indoor Environment Based On LIDAR Assisted By MEMS Inertial Components

Simultaneous Localization and Mapping is also called Concurrent Mapping and Localizationwhich can provide robot real-time surrounding unknown environmental survey and relativepositioning, it is a necessary system for robot to achieve unknown environmental survey andautonomous navigation positioning. It will expand the application scope of SLAM system greatlywhen SLAM system is applied to the unmanned aerial vehicle(UAV)autonomous navigation andenvironmental survey, Meanwhile, it will enhance the independent ability of UAV greatly which is thekey nodes of UAV to a higher level of autonomous ability development and an important part of theUAV to artificial intelligence. This dissertation mainly aims at SLAM system based on UAV platformfor research, analyzes limiting condition of SLAM algorithm applied to the uav platform, studies theKey technology of MEMS data integration、feature detection、feature matching、the compensation ofenvironmental characteristics when attitude changed.This dissertation studies the preprocessing algorithm of MEMS aided laser radar sensor signal,puts forward the acquisition scheme of attitude data outputted by MEMS and LADAR. To extract thefeature model in the environment, this dissertation studies the algorithm of corner and line featureextraction, which improve the feature extraction accuracy effectively.After that, to restrain MEMS’s excursion, meanwhile, reduce the effect of attitude algorithmwhen vehicle speeding up, This dissertation studies the complementary filter algorithm of attitudeoutputted by ultrasonic array attitude measurement system and MEMS, calculates attitude、positionincrement、 speed increment with the use of data outputted by MIMU,MEMS data conductscomplementary filter with the attitude measured by ultrasonic array attitude measurement system, usethe attitude after filtering to assist LADAR compensate distance-measuring error. Restrain MEMS’sexcursion effectively, at the same time, provide LADAR precise attitude information.To deal with the errors in radar parameter extraction caused by the particularity of attitudechange in UAV flight, an algorithm is proposed to compensate for the characteristic parameters of theenvironment when attitude changes. The global map is established based on the local feature map.The double feature library screening method which screens feature points in the local map and mixesmatching features of trusted corner feature and line features with features of the library is proposed.Updates of the global map and UAV indoor self-positioning in unknown environment is achieved based on the research.At last, this dissertation sets up a experimental platform of LADAR Simultaneous Localizationand Mapping system assisted by MEMS, carries out flight experiment in indoor environment whichprovide the flight result verifying the feasibility of system program.