Home > Research > Publications & Outputs > A robust evolving cloud-based controller


Text available via DOI:


View graph of relations

A robust evolving cloud-based controller

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNChapter (peer-reviewed)peer-review

Publication date2015
Host publicationSpringer Handbook of Computational Intelligence
EditorsJanusz Kacprzyk, Witold Pedrycz
Place of PublicationBerlin
Number of pages15
ISBN (electronic)9783662435052
ISBN (print)9783662435045
<mark>Original language</mark>English


In this chapter a novel online self-evolving cloud-based controller, called Robust Evolving Cloud-based Controller (RECCo ) is introduced. This type of controller has a parameter-free antecedent (IF) part, a locally valid PID consequent part, and a center-of-gravity based defuzzification. A first-order learning method is applied to consequent parameters and reference model adaptive control is used locally in the ANYA type fuzzy rule-based system. An illustrative example is provided mainly for a proof of concept. The proposed controller can start with no pre-defined fuzzy rules and does not need to pre-define the range of the output, number of rules, membership functions, or connectives such as AND, OR. This RECCo controller learns autonomously from its own actions while controlling the plant. It does not use any off-line pre-training or explicit models (e. g. in the form of differential equations) of the plant. It has been demonstrated that it is possible to fully autonomously and in an unsupervised manner (based only on the data density and selecting representative prototypes/focal points from the control hypersurface acting as a data space) generate and self-tune/learn a non-linear controller structure and evolve it in online mode. Moreover, the results demonstrate that this autonomous controller has no parameters in the antecedent part and surpasses both traditional PID controllers being a non-linear, fuzzy combination of locally valid PID controllers, as well as traditional fuzzy (Mamdani and Takagi–Sugeno) type controllers by their lean structure and higher performance, lack of membership functions, antecedent parameters, and because they do not need off-line tuning.