Research On Map Building For Multi-Robot System

Mobile robots which can autonomously accomplish tasks in unknown environment have wide potentials for application in industry, civilian life and military. Autonomous mobile robots should be able to explore and map unknown environment for accomplishing various tasks. Map building embodies the perception ability and intelligence of robots. The dissertation focuses on the related problems of map building in unknown environment. The main work of this dissertation is as follows:Firstly, the dissertation reviews the state-of-the-art research of map building of mobile robot and introduces background and main contents of this thesis.Secondly, D-S evidence theory is applied to fuse the sonar data for occupancy grid map building in indoor structural environment. Self adaptive RCF factor is used to eliminate uncertainty. We have done experiments with the Pioneer3 mobile robot. Experiment results prove that data fusion algorithm based on D-S evidence theory gain satisfactory effect in modeling environment for mobile robot.Thirdly, hough transform which is usually used in image process problem is applied to extract the line characteristic of original sonar data for eliminating the impact of uncertain information further. After sonar data being preprocessed, not only the quality of the built map but also mapping efficiency is greatly improved. This will provide good basis for fast and efficient map building of mobile robot in unknown environment and improve the ability of autonomous operation, navigation and path planning further.Fourthly, for improving the efficiency of implementing expected tasks, we have done some research work on map building for multi-robot. After reviewing and analyzing the algorithm of map merging for multi-robot, the paper implements the map merging process for mobile robots without knowing the relative positions of robots. Based on the map similarity concept, this dissertation applies the genetic algorithm search strategy to search in the global world. The parallel search strategy improves the efficiency of globally search and effect of map merging.Finally, conclusions are provided. Outlook of future research is ananlyzed.