Study on robot perception system of multi-sensors information fusion based on fuzzy neural network
Lanshen Guo, Fang Wang, Minglu Zhang, Lixin Sun · 2008
This paper developed three subsystems of robot perceptive system: visual, aural and olfactory for a three DOF robot perceptive system, which can get the primary location of the target by aural sensor, the exact location and tailing of the target, and estimation the location parameter by CCD camera. The olfactory sensor was used to detect whether there were dangerous gases around or not. The multi-sensor fusion model and arithmetic based on fuzzy neural network were presented in this paper, design the input and output of every layer, utilized BP arithmetic to adjust the weight and the parameter of the fuzzy neural network and obtained different subject functions thereby created relevant fuzzy rules to actualize fuzzy decision. Finish the precise location and real-time tracking. Simulating the perceptive process of the robot used Matlab, obtained the curves of the practical output and the experiment output, validated the validity of using fuzzy neural network to fuse the visual and aural heterogeneous-sensors.