It is a well known fact that sliding mode control (SMC) is a powerful control method ology for both linear and nonlinear systems because of its robustness to parameter changes, external disturbances and unmodelled dynamics. Besides its power, the design of sliding mode controllers needs the information of the system's state, which makes the design relatively austere in some applications where the mathematical modelling of the system is very hard and where the system has a large range of parameter variations together with unexpected and sudden external disturbances. For those applications, a controller that will provide predicted performance even if the model of the system is not very well known, is needed. That controller should also adapt itself to large parameter variations and to unexpected external disturbances. These types of controllers are generally called 'intelligent' controllers, mainly work ing on the principles of fuzzy logic, neural networks, genetic algorithms and other technologies derived from artificial intelligence. The idea of combining these intelli gent control structures with the power of sliding mode control approach has attracted much research. A recent survey on the combination of SMC and intelligent control can be found in Reference 1. In this chapter, the union of sliding mode with neural networks and fuzzy logic is examined with examples from literature, and then a new technique combining neural networks and sliding mode control is presented.

Inspec keywords: variable structure systems; fuzzy control; neural nets

Other keywords: neural network systems; fuzzy control; parameter variations; mathematical modelling; fuzzy logic; genetic algorithms; SMC; sliding mode control

Subjects: Multivariable control systems