Hierarchical control structure for multisensor object recognition system

Multisensors can be used to overcome problems associated with intelligent object recognition systems. The introduction of multisensors into such a system emphasizes the need for an effective method for combining sensor outputs, and a complicated system requires intelligent control. We propose a hierarchical control structure (HCS) to build a recognition system and introduce a fuzzy neural net (FNN) method to fuse multisensor information. This paper describes the concept of HCS, the derivation of fuzzy rules, the FNN's construction and training, etc. Then, experimental results involving parts recognition in the application of an assembly robot are presented.<>