Microcontroladores: definição e implementação de controladores estruturalmente flexíveis para sistemas adaptativos

Abstract

Context: This work starts from the hypothesis that controllers for adaptive systems that are partitioned into elements, in this work called microcontrollers, which are dedicated to each service involved in the stages of a control loop have the advantage of structural flexibility without compromising the reconfiguration and performance of the target system. Objective: This work aims to propose and evaluate an approach to develop controllers that are structurally flexible (i.e. based on microcontrollers), investigating advantages of the approach by comparing with different configurations of controllers. Methodology: In order to achieve the objective, the following activities were performed: (1) definition of flexible controllers by promoting them as a set of independent microcontrollers, demonstrating them as microservices implementations of the control loops; (2) conduction of an exploratory study and associated qualitative analysis regarding the reconfigurations of the controller; (3) execution of comparative studies to evaluate the approach that is being proposed against other approaches found in the literature; and (4) specification of an implementation by using the microcontroller-based approach, to demonstrate the flexibility of the approach. Results: the results and contributions are: an approach to define controllers that are structurally flexible; an exploratory study involving a qualitative analyse about reconfiguration in design time and runtime; a set of comparative studies evaluating different approaches from the literature; a proposal involving the evolution of the multi-layered controller to demonstrate the feasibility of the approach of this work. Conclusion: As conclusion, a multi-layered controller design, based on micro-controllers, provides the basis for defining structurally flexible controllers at operational-time, and may promote reuse at development-time. In addition, even demanding more computational resources, the approach does not impact on the performance and on the reconfiguration of the target system.