Research On The Problem Of Simultaneous Localization And Mapping For Multiple Mobile Robot

With the development and maturation of the technology of mobile robot, the applicationsand demands of mobile robot are also increasing. The research level and applications of mobilerobot reveal the development of the industrial automation of a country, thus they are significantto the national defense and strategy. Simultaneous localization and mapping (SLAM) refers tothat the mobile robot builds the environment maps with the equipped sensors, and estimate theirpose simultaneously. As it’s the key issue to achieve autonomous mobile robot, it’s also adifficult problem in the field of robotics, incarnating a concentrated reflection of mobile robotperception ability and intelligence. In addition, with the expanded application of mobile robot,single mobile robot often has a difficulty to cope with the complex environment and tasks.Compared to a single robot, multiple robot cooperative simultaneous localization and mappingnot only improves the efficiency of the composition, but also increases the efficiency androbustness of the whole system. Thus, the technology of cooperative multiple robotsimultaneous localization and mapping has become one of the most hottest issues in the field ofmultiple robot system. This paper studies the state estimation, data association, and map fusionproblem related to multiple robot simultaneous localization and mapping. The main works ofthis thesis are as follows:(1) Firstly, research on the overview construction of the multi-robot system. Threecommon used system architecture was introduced. And then discussed the map describe methodand established the model of mobile robot. Secondly, the state estimation problem of singlemobile robot was researched. Research starts from Kalman filter, then comes to extendedKalman filter. After that, using extended Kalman filtering technique to achieve the position andorientation of the robot synchronized estimated. And an extended Kalman filter based SLAMalgorithm and simulation systems are realized using MATLAB.(2) The data association problem of SLAM was researched. Aim at the problem that the accuracy and speed of the traditional association algorithm decreased when it comes to theapplication in large-scale environments. This paper proposed an improved joint compatibility andbranch and bound algorithm. It has a good efficiency and high speed, which guarantees thereal-time and robustness of the SLAM algorithm.(3) The problem of collaborative multiple robot was studied. According to whether two robotswill meet or not, two map fusion methods was approved. The first one uses relative observing modelto get the relation matrix of two sub-map in order to map fusion. The second one transforms themultiple map fusion problem into objective optimization problem, and then use particle swarmalgorithm to find the best transformation matrix between two sub-map to realize map fusion. Twomethods are used according to proper situation so that collaborative SLAM is achieved.At the end of the paper, some conclusions of the research work are made and some outlook ofthe next research direction or problems are discussed briefly.