SAFEST

A Digital Twin (DT) is a machine-processable high-fidelity virtual representation of a physical system, called Physical Twin (PT), to which it is coupled through a continuous, bidirectional flow of data (e.g., monitored data and results of predictive/prescriptive analysis). The DT paradigm has emerged as a suitable way to cope with the complexity of analyzing, designing, implementing, controlling, and adapting complex systems belonging to diverse domains such as cyber-physical, business, and societal systems. In the context of Medical Cyber-Physical Systems (MCPSs), which are critical, interconnected, distributed and context-aware systems of devices used in medicine, DTs can simulate physical devices and specific machinery pieces to provide more informed and real-time relevant healthcare responses, enable decision making and assess risks for the patients; training medical staff is a further highly required application of DTs for MCPSs. However, the DT paradigm is not fully implemented in the medical setting. The complexity of the modeling of human behaviors and workflows, the presence of uncertainty, and the need for security guarantees for sensitive information (e.g., patient personal medical profiles) are barriers to the adoption of DTs. In this context, the SAFEST (truSt Assurance) project aims at improving the application and soundness of DT-based of digital twins For mEdical cyber-phySical sysTems methodologies and tools. To win this challenge, SAFEST identifies and focuses on two main goals:

• taming the complexity caused by the heterogeneity of the DT components that must be built, operated, coordinated, and evolved together with their physical and human counterparts;
• increasing the level of trust in the results and indications coming from a DT, despite modeling approximations and uncertainties caused by incomplete or imprecise data collected in the field.

SAFEST intends to go beyond the usual interpretation of trust, commonly defined in terms of security and privacy, and takes into account a broader set of dimensions, including performance, dependability, safety, and conformance to required behavior. SAFEST will pursue these two goals by articulating them in the following two objectives:

• O1: Modeling notations for evolving heterogeneous systems with uncertainty (multi-view, multi-paradigm, multi-dimensional, O1 multi-disciplinary models);
• O2: Trust assurances (behavioral conformance, safety, dependability, security, performance). O2 The suitability and effectiveness of the proposed methods and tools will be evaluated through a case study in the medical domain.

Scientific goals of the SAFEST project are in line with the following missions of the Italian national plan on Recovery and Resilience (PNRR), more precisely mission 1.2 “Digitisation, innovation and competitiveness of the production system”, mission 6.2 “Innovation, research and digitisation of healthcare”, and mission 4.2 “From research to business”.

Partners

University of Bergamo: Angelo Gargantini (PI), Patrizia Scandurra, Silvia Bonfanti

University of Milano

Politecnico di Milano

Diapath