A robot self-localization method based on perception dominance

As fixed robot localization method is not well suitable to different system, a self-localization method merging with dominating competition mechanism is proposed. Firstly, based on the combination of multi-dimensional perception, a self-organizing perception system with unified interface is built. Then, evaluation vector can be established according to evaluation principles of each perception group. The concept of “dominance” is introduced to evaluate the dominating degree of different localization methods through self- localization. The proposed system can timely choose the optimal positioning method which is simple and robust by D-S evidence theory even when the environmental changes, failure and recovery degree of positioning sensor and reliability changes. This method behaves well in validity, practicability and robustness.