DSmT-based Generalized Fusion Machine for Information Fusion in Robot Map Building

摘要

Characteristics of uncertainty and imprecision, even imperfection is presented from knowledge acquisition in map reconstruction for autonomous mobile robots navigation. A method is presented for building grid map using sonar measurements. A generalized fusion machine including an ESMS (Evidence Supporting Measure of Similarity) information filter, a fusion operator based on DSmT (Dezert-Smarandache Theory) and a conflict redistributor based on PCR5 (Proportional Conflict Redistribution No.5) rule is proposed. It is applied to map-building of mobile robot with the help of selflocalization method based on δ neighboring field appearance matching arithmetic. The general basic belief assignment (gbba) functions are also constructed according to sonar's uncertainty. An experiment using a Pioneer Ⅱ mobile robot with 16 sonar detectors onboard is done in a small indoor environment, and a 2D Map is built online with our self-developing software platform. A comparison of the results of our new method for map reconstruction with respect to those obtained from classical ones is provided,and how the new tool proposed in this work outperforms other approaches is shown. This study provides aside an useful human-computer interface for a mobile robot exploring unknown environment, and for path planning and navigation. It also establishes a firm foundation for the deep study of information fusion theory.