Adaptive Supervisory Framework for Cyber-Physical Systems

Authors

Abstract

Cyber-Physical Systems (CPS) result from the aggregation of the cyber and physical worlds into a single spatially distributed macro-system, relying on sensor and actuator networks to perform its optimised management.

CPS must be capable of reacting to perceived and inferred knowledge about the systems they coordinate, demonstrating context-aware and self-adaptive characteristics. Popular solutions proposed in the literature are inspired by feedback-loop control theory, according to which a repeated analysis of the state of the managed environment is performed, followed by a review and adjustment of operational objectives and constraints.

Despite current advances CPS still face important challenges, among which the integration of heterogeneous entities (where human presence must also be accounted for), high-order predictive models, lack of a holistic design perspective and inclusion of context information are highlighted.

The proposed research aims to contribute towards a clarification of these gaps by addressing the design of CPS, integration of dynamic and heterogeneous entities, and improving their overall optimised behaviour under changes and uncertainties in perceived and inferred context information. To this end an Adaptive Supervisory Framework is intended to be developed.

Keywords

Cyber-Physical Systems, Context-Awareness, Human-in-the-Loop, Internet of Things, Supervision

TechReport Number

TR 2018-001

PDF File