Autonomous Learning Controllers
Plamen Parvanov Angelov · 2012
This chapter discusses the principles of indirect adaptive control scheme. It presents a brief description of the inverse plant dynamics model and also discusses the autonomous controller structure from online streaming data. Indirect adaptive control is a scheme that aims to substitute the inverse plant dynamic by a model. One way to train the controller is to use offline learning and apply neural networks (NN) or a fuzzy rule-based (FRB) system because both are proven universal approximators. AutoControl can start from a very simple configuration or even ‘from scratch’ and self-develop its structure based on the data obtained online. The chapter shows two simple examples with AutoControl primarily with illustrative purposes. Controlled Vocabulary Terms adaptive control; fuzzy control; learning (artificial intelligence); Streaming media