Fuzzy Cognitive Map Based Situation Assessment Framework for Navigation Goal Detection

Navigation goal detection in complex real world scenarios is challenging for autonomous mobile robots due to the uncertainty and ambiguity of their environments. This paper proposes a fuzzy cognitive map (FCM) based situation assessment framework (SAF) for mobile robot navigation goal detection. A given navigation goal is described with several sub-goals using prior knowledge. Based upon these sub-goals, the proposed SAF operates recursively on the navigation goal to verify it. The decision fusion system combines sensory information from multiple sensors to verify the sub-goals. The FCM, which is realized using fuzzy gamma fusion, is used as a high level reasoning engine. The navigation goals are verified based on the rules which connect sub-goals together and the goal assertion confidence of sub-goals which is decided using FCM inference. Experimental results demonstrate that the proposed framework can accurately decide the navigation goals in unknown environments based on the sensory information and expert knowledge.