GMP Based Fuzzy Reasoning: An Application to Sonar Based Navigation
Kudret Demirli*, Burhan Türksen**, and Mohammad Molhim*
*Fuzzy Systems Research Laboratory, Department of Mechanical & Industrial Engineering, Concordia University, Montreal, Quebec, Canada, H3G 1M8
**Department of Mechanical & Industrial Engineering, University of Toronto, Toronto, Ontario, Canada M5S 3G8
Generalized Modus Ponens (GMP) type fuzzy reasoning is used for mobile robot navigation in this paper. However, GMP type reasoning suffers from different drawbacks that make it difficult to be used in real time applications. The objectives of this paper are to eliminate some of the drawbacks of GMP type fuzzy reasoning to make it favorable in control applications, and also demonstrate the effectiveness of the suggested improvements on a real-time application. The first drawback of GMP is its high computational requirements and the second problem is the lack of a systematic operator selection procedure. There are also other issues to be dealt with when the conclusions obtained from GMP are to be combined and defuzzified in order to be used in the control of an electro mechanical device. These drawbacks are tackled in appropriate ways and it is experimentally shown that GMP can be satisfactorily used for mobile robot navigation.