Neural networks and fuzzy logic in active control of mechanical systems

A few applications of neural networks and fuzzy logic in active control of systems are presented. Rigid and flexible, linear and nonlinear, stable and unstable structures are investigated. The basics of neural networks are not covered unlike the essentials about fuzzy reasoning that are highlighted. Within the presented control algorithms neural networks accomplish different tasks. They cover cases in which the control action is computed according to a neural-only paradigm all the way through simpler applications where the role of neural networks is to replicate the behavior of conventional controllers. Nonlinear civil structures under seismic excitation and nonlinear rigid systems as an inverted pendulum and a container-ship are among the systems studied in much detail. Attention is focused on the physics of the problem, on the objectives of the neurcontroller as well as on its practical implementation. Numerical simulations and experimental tests are illustrated to validate the presented approaches.