A Sensor Fusion Technique Utilizing a Genetic Approach for Networked Sensing Systems - Application to a Recognition System for Object Shapes

This paper describes a networked sensing system constructed by a novel sensor fusion technique with a genetic approach. Conventional sensor fusion approaches do not function well when they are applied to situations which the designers of the system have not anticipated. The proposed sensor fusion system aims to apply to unknown situations by implementing GA. This paper examined a prototype object recognition system as a concrete example of the networked sensing system. From the experimental results, it is shown that the proposed networked sensing system is effective for the many applications