Research Of Self-localization Technology For Intelligent Wheelchair Under Indoor Enviroment

Since self-localization has been hot and difficult problems of mobile robotics research, itis the same to indoor intelligent wheelchair. It is the primary solution for autonomousnavigation of the intelligent wheelchair. In particular, the wheelchair starts from any position,it can acquire its current position in real time, and can recover itself from lost state during theprocess of moving. In order to achieve the self-localization of the intelligent wheelchair in theindoor environment, the research work of this paper is closely around the key technologiesand difficult points of intelligent wheelchair self-localization. They are expanded from theenvironment perception and description, as well as self-localization algorithm. We havedeveloped a set of localization software platform and conducted in-depth research ofself-localization technology for intelligent wheelchair under indoor enviroment, with theintelligent wheelchair served as experiment object.First of all, we have conducted a detailed introduction and overview of intelligentwheelchair localization technology, respectively from the aspects of the coordinate systemmodel, environment description method, localization method and experiment platform.Intelligent wheelchair relative localization method based on encoders was studied from theaspects of theory and experiment. It turned out that the encoders used in this paper apply toclose range and high precision localization and it should be combined with other localizationmethod to achieve remote localization.Secondly, according to the characteristics of the indoor structured environment, andthrough the analysis and comparison, the vector sequence map method to create a global mapis put forward, in order to improve the efficiency and accuracy of the matching algorithm. TheSICK laser radar LMS111as wheelchair distance sensor was selected and the least squaresstraight fast vector extraction method was used to extract vector line from laser radar scanningdata. At the same time, the creation of a local map was completed by the vector sequence mapmethod.Finally, taking into account the initial localization and the process of localization bothneed with the help of a laser radar for local environmental information to implement thelocalization, vector line segment relationship matching method was put forward in this paper, by establishing a corresponding relationship matrix to record the vector line segmentcorresponding relations of local map and global map, at the same time, the improved bestmatch search method was proposed to search for the maximum value in the relationshipmatrix. After getting the best match vector line segment pairs to estimate the position andorientation of the wheelchair, the different implementation principles of initial localizationand process localization were given in detail respectively, and the experiments on the vectorline segment relationship matching method were tested.Through these series of detailed work and a number of proposed improved algorithms,the self-localization method of intelligent wheelchair under indoor environment is feasibleand effective and has a high accuracy and efficiency, as well as good practical value.