Supervisor(s): Dr W Naeem

For several years, unmanned systems such as industrial robots are being effectively utilised in a structured environment for instance, in manufacturing applications. Now there is a growing interest in developing mobile platforms, with improved intelligence, which are able to operate in any unknown environment. Also the dynamics of a system could subject to changes such as sensor faults or even damage to the onboard components. In these types of scenarios, the objective is to re-establish the structure of the system with a minimum shutdown time. The main target is thus to design a control system which can recognise and respond to changes and failures in the dynamics of the system. In addition, it is common for researchers to spend an extortionate amount of time to develop accurate mathematical models. These models are employed to predict the system's response precisely and for devising new control strategies.

These issues are hard to realise with traditional modelling and control schemes, hence, it would be necessary to incorporate nonlinear adaptive control methodologies where machine learning algorithms, neural networks and reinforcement learning can achieve a dynamic adaption in response to system changes.

This project will investigate advanced self-modelling and adaptive control algorithms for unmanned systems such as robots with a special focus on improving the overall development time.