Mobile Robot Instant Indoor Map Building And Autonomous Navigation

Mobile robot autonomous navigation is a hotspot in the field of mobile robot problems. And service robot is an important classification of mobile robot. A service robot not only needs to grasp the information about environment, but also to locate itself. This paper mainly studies the mobile robot instant indoor map building and autonomous navigation.This paper presents the situation and hot trends about mobile robots first, analyse and take some different robot platforms and sensors for comparison. Introduce Pioneer-3 DX type mobile robot based on ARIA software system briefly, also state the robot C/S communication structure and the advantages and disadvantages that mobile robot autonomous navigation based on different map-building methods.This paper presents a method that mobile robot build a topological map and simple localization based on the characteristics of nodes. Mark the main nodes also namely room that can be extended. Build an instant map after the robot enters a room. First, observe the room and select the center in the open area, adjust the direction of the robot and move to the center. Then, scan the room from the center using MbICP method to match each group of scan data points that the robot gets and using grid to divide the data points. At last, using weighted least squares fitting to optimize the grid points. We can extract a prior map that can be able to represent the whole outline of the environment. After that, the mobile robot can match each cycle's data points with the prior map points using MbICP and take the results as an observation. Then, we use EKF method to localization. The method that we present comprehensive the good advantages of MbICP algorithm and EKF filter algorithm.The experiment in this paper based on Pioneer-3DX type mobile robot's platform and LMS200 type laser ranger. Acquire a series of experimental results and data through programming the instant map building and autonomous navigation system. After analyzing and verifying the results and the data, prove the method that this paper presents is effective, robust and practical.