Generalised Controller Design Using Continual Learning

In control systems applications, controllers for different plants are usually designed with different methods. Although plants may share common characteristics, these controllers are generally designed in isolation. The problem of continually learning a sequence of related tasks has been extensively studied recently. A challenge in continual learning is the phenomenon of catastrophic forgetting of knowledge of previous tasks which have been integrated into a neural network model. In this paper we evaluate the feasibility of modelling different controllers using continual learning. We explore regression versions of state-of-the-art methods and demonstrate that even the simplest continual learning approach decreases the overall Mean Average Error (MAE) by 39% of the MAE achieved by a non-continual strategy. Furthermore, a method based on dynamically expanding the network can achieve an overall MAE which is only 18% of the non-continual MAE. We also propose a set of new metrics that allow us to characterise the nature of catastrophic forgetting experienced while using different continual learning methods.