Brown, M., Fraser, R.J.C, Harris, C.J. and Moore, C.G.
Intelligent Self-Organising Controllers for Autonomous Guided Vehicles: Comparative Aspects of Fuzzy Logic and Neural Nets.
Full text not available from this repository.
Of central importance to autonomous guided vehicles (AGVs) are the adaptive or intelligent tasks of:- (i) Multi-sensor data fusion or integration of sensory data for vehicle location, to represent or to model its internal states and its environment. (ii) Planning and navigation. (iii) Motion control. This paper addresses the last issue, motion controls that are vehicle dependent. Classical motion control is based upon deriving a set of vehicle model equations, and synthesising a set of feedback control laws. Unfortunately such an approach is limited only to constant velocity and environmental conditions and for small perturbations. Yet humans are able to generate driving algorithms with little physical insight, but with a great deal of experimental knowledge, this is flexible, robust, sufficiently precise for proper functioning, and intelligent, in that they adapt to differing environmental and payload conditions. This paper will review two approaches to AGV intelligent control adopted at Southampton, - (i) Self-organising fuzzy logic controllers (ii) associative memory type neural nets. Both approaches have the potential to provide real time adaptive convergent, robust decision strategies with little apriori knowledge. There are many similarities between the two approaches: (i) a transformed input space is required in both cases, (ii) Initial, approximate plant models are required, (iii) both adopt local weight/rule adaptation schemes - we also show an equivalence of learning through LMS based adaptation rules based upon B-splines, (v) both methods are extremely robust and fault tolerant.
Actions (login required)