Algoirthm Research On Multi-sensor Information Fusion Based On D-S Evidence Theory

Because of its simple structure, high system stability and wide application, Mobile robotshave been a hot research. Mobile robots’ ability to create a map in an unknown environmentis an important topic in the field of mobile robot research. It is the basis on autonomousnavigation of mobile robot and completing complex tasks. Because of the complexity and theunknown of the mobile robot’s working environment, the information provided by a singlesensor is generally one-sided and fuzzy. So we must rely on multi-sensor information fusiontechnology for mobile robot research. This paper mainly studies algorithm research on multi-sensor information fusion based on D-S evidence theory and its application of mobile robotmap building. The main research works are as follows:Firstly, this paper introduces the sensor classification of mobile robot system and thecommon map representation methods. This paper introduced the research status of mobilerobot mapping based on multi-sensor information fusion and introduces the background andmain research contents of this thesis.Then, in order to create a grid map, this paper defines the coordinate system of themobile robot and grid map representation. At the same time, we used the ultrasonic sensor andlaser range finder in this paper. This paper introduces ranging principles and the observationmodels of two kinds of sensors. Based on the above basis, the article describes the processand method of building grip map based on D-S evidence theory.Secondly, the sensors will get a lot of uncertain information.This paper introduced theHough transform to preprocess the sensor data, which removed a lot of uncertain information,reduced the workload of D-S information fusion greatly and improved the efficiency of mapbuilding. Because there are some limitations in combination rules and decision rules of D-Sevidence theory. We improved the combination rules and decision rules of D-S evidencetheory which is suitable for grid map building. And given the creating process of raster mapimprovement based on D-S evidence theory, carried on simulation experiment and verifiedthe effectiveness and feasibility of the method.Finally, we carried out experiments based on the ability storm robot platform. Theexperimental results shows that the multi-sensor mobile robot grid map building methodbased on improved D-S evidence theory is appropriate.