The Control Based on Lyapunov Adaptation Law to be Improved by Behavioral Cloning and Machine Learning

Development of adaptive control system for over its history has a great number of approaches. A crucial property of all approaches is that they have the ability to adapt the controller to accommodate changes in the process. That means that the controller maintains a required level of performance in spite of noise and fluctuation in the process. The paper focuses on model reference adaptive controllers based on Lyapunov adaptive law. The named adaptive controller in some cases has a little level of performance regarding conventional controller. In order to improve the controller performance behavioral cloning is used. In the behavioral cloning, the system learns from control traces of a human operator. The advantage of the named approach lies in simplicity with regard to the controller parameters tune up in order to reach an appropriate performance.