Abstract: Multi-source information fusion algorithms were used in the searching of gas leakage source in this paper.In order to improve the efficiency,gas sensors and vision sensor were mixed to get information,single gas sensor measurement was replaced by the multiple gas sensors,and the measurement method was changed from single-point to multipoint.In each data fusion period,some suitable algorithms are applied to determine the searching direction of the robot.The result indicates that the weighted average method used to fuse the data of gas sensors with same kind can reduce the influence of the noises and other troubles,the least square method which can optimally estimate unknown parameters can originally estimate the location and flow of the leakage,and the Bayesian inference can accumulate different information source to judge the leakage,leading the robot to search the target more reasonable.

Key words: Multiple-source information,Information fusion,Harmful gas,Leakage source,Searching