Mobile Robot Slam Problem

Mobile robot has vast application prospects in areas including space exploring, factory automation, mining, eliminating dangerous situation, military and service. SLAM (Simultaneous Localization and Mapping) means mobile robot create environment map autonomously in unknown area, and localizing itself in this map simultaneously. In this process, the robot has no help of prior knowledge of the environment, and achieving localization and mapping with its onboard sensors. Thus, the solution of SLAM problem is the key to mobile robot automation.This paper is focused on solving SLAM problem in unknown outdoor environment. First we summarized the research of SLAM problem, analyzed the pros and cons of existing research. In order to solve robot "kidnapped" problem, we proposed a new solution to SLAM problem. The main contribution of this work is:1. New landmark detection algorithm. The basis of modeling the environment and scan matching is the landmarks in the environment. In order to extract landmark points from laser range data, we proposed a new entity landmark detection method. The spider web is introduced in to separate objects that occupy a range in the environment. The entities which meet certain standards are extracted as landmark points.2. New SLAM problem solution framework. The algorithm is based on scan matching between local map and global map. Data association matches landmarks in local map and global map, and generates a match matrix. The core of the method is multiple hypotheses tracking, which improves the robustness of the SLAM solution.3. Introducing Delaunay triangulation into map generation. Based on landmarks extracted from the environment, we use Delaunay triangulation to construct an optimized triangular mesh as the map.4. Introducing in Singular Value Decomposition to calculate robot poses. The accurate robot poses are calculated based on the matched landmark pairs by SVD.We compared the proposed method to traditional algorithms; the experiment results prove the validity and advantage of the proposed method.