A Model-Driven Approach to Automate Tuning of Continuous Controller Parameters

Cyber-physical systems evolving in uncertain environment endure fluctuating dynamics during their lifetime. Variations can be related to environment evolution, physical damages and component failures occurring after deployment. In such a variable context, controlling systems towards safety and system performances is challenging. In particular, controller definition and tuning (finding optimal control parameters) are key points of the development process. Determining optimal control parameters and the boundaries of a controller is a challenging process due to the multiplicity of contexts to be considered during the tuning phase. The challenge is here to identify good control parameters for the different contexts, considering multiple variation points. In this paper, we use a combination of model-driven simulation and dimensionality reduction techniques to define adequate control settings, considering multidimensional inputs defined by hardware and environmental parameters (the context). First, we define criteria of evaluation based on user quality of control requirements. Then, we apply feature selection algorithms to identify contextual parameters that have a significant impact on performances of the controller. Considering only selected parameters, we finally carry out a clustering aiming at identifying for context domains an optimal control parameters setting. To evaluate the proposed approach, we apply it to a proportional controller used in the context of a leader/follower application. The experiment shows effectiveness in the identification of optimal control parameters setting of the controller, for different contexts. The obtained results are then used to adapt the values of the controller parameters depending of environmental context.