Neurocontrol and fuzzy logic: Connections and designs

Artificial neural networks (ANNs) and f u z z y logic are complementary technologies. A N N s extract information f rom systems to be learned or controlled, while fuzzy techniques most often use verbal information f rom experts. Ideally, the two sources o f information should be combined. For example, one can learn rules in a hybrid fashion and then calibrate them f o r better whole-system performance. A N N s offer universal approximation theorems, pedagogical advantages, very high throughput hardware, and links to neurophysiology. Neurocon t ro l the use o f A N N s to directly control motors, actuators, e t c . uses five generalized designs, related to control theory, that can work on f u z z y logic systems as well as ANNs. These designs can copy what experts do instead o f what they say, learn to track trajectories, generalize adaptive control, and maximize performance or minimize cost over time, even in noisy environments. Design trade-offs and future directions are discussed throughout. The final section mentions a f e w new ideas regarding reasoning, planning, and chunking, with biological parallels.