| How to localize mobile robot effectively is a challenging research topic, and with the further development of robot technology, its importance will increasingly highlight. This paper is mainly to explore the method of indoor localization using multi-sensor informa-tion fusion in the known structured enviroment. We propose a thinking about introducing redundant parameters into the measurement equation of Kalman filter algorithm to im-prove the precision and stability of indoor positioning. The main contents include the following sections.First, biulding mobile robots for research and validation algorithm, and equipped with various sensors required for the experiment.Writing low machines and PC programs about controlling mobile robots, sensor data acquisition and display.Second, modeling and error analysis for encoder and laser rangefinder sensors, these are the important basis for selection of information fusion algorithm and determing of system parameters later.Third, for the characteristics of data collected by the laser rangefinder, using a adap-tive threshold algorithnm for data clustering. According to our assumptions about the known environment and choose the parameters can be reasonably characterize the envi-ronment, using the Iterative End Point Fit method to extract corner features of high re-liability from the data. Comparing the mathmatical model of the corner features extract from raw data with parameters of corner features of offline map to complete feature points matching.At last,Proposes the idea of the introduction of redundancy parameters in the con-ventional Kalman measurement equation in order to improve the effect of indoor position-ing. Simulation and experimental verification algorithm. For update section of unscented Kalman filter, introducing redundant information of corner points which based on polar coordinates and Cartesian coordinates for indoor localization. And performing relevant comparative experiments and doing analysis. |
PDF Full Text Request