Research On Localization Technology For Indoor Autonomous Mobile Robot

Localization is the basic function and the critical step towards autonomous navigation of autonomous mobile robot. It is of great significance to improve the automatization level. In this dissertation, a deep analysis and systematic study of technologies related to the mobile robot's absolute positioning are carried on.Firstly, a wireless sensor networks is introduced into the research of mobile robot. This paper presents a localization method based on the ZigBee wireless sensor networks. This method is similar to a small global positioning system (GPS), and it can be used in place of satellite systems for mobile robot localization while which can not or should not be used.Secondly, on the basis of review and summary of methods on mobile robot environmental model, this paper proposes a new geometry map building approach based on two-dimensional laser range sensor, namely, line segment and feature point are adopted to describe environment. The method of environmental feature extraction is described in detail in order to build a relatively complete map of the environment.Thirdly, this paper presents a localization method on the basis of complete line segments matching when the environmental model is established. The paper establishes match assumptions based on the length of complete line segments, and evaluates the match assumptions using the relative relationships between complete line segments and feature points in order to establish the match matrix and the flag matrix. The robot pose is obtained based on the best match which is searched according to the match matrix and flag matrix using the best match search algorithm.Finally, the paper presents a rapid localization method for an indoor mobile robot using the integrated information from the ZigBee wireless sensor networks and laser range sensor. The algorithm achieves a rough global location of the mobile robot by using the ZigBee wireless sensor networks, and extracts the reference map close to the mobile robot from the known global map. Then, the algorithm matches the known global map and the environmental features extracted from the laser range sensor information. According to the matched environmental characteristics, the accurate position of the robot can be obtained.