Context and Motivation:
Systems that combine digital and analog devices, interfaces, networks, computer systems, and the like, with the natural and man-made physical world are called Cyber-Physical Systems (CPSs). The inherent interconnected and heterogeneous combination of behaviors in these systems makes their analysis and design a challenging task. Safety and reliability specifications imposed in cyber-physical applications, which are typically translated into stringent robustness standards, aggravate the matter. Unfortunately, classical tools for system analysis and design cannot cope with the intrinsic complexity in CPSs. Tools suitable for analysis and design of such systems must allow a combination of physical/continuous dynamics and the cyber/computational components comprising the networked structure, as well as handle a variety of types of perturbations, such as exogenous disturbances, time delays, computation effects, and system failures.
Tools for that purpose are becoming available in the literature, but most algorithmic designs nowadays are still based on ad-hoc solutions trying to make the connection between the classical techniques of systems and control theory with those in software systems.
Most research in CPSs considers design of algorithms and their implementation separately. Algorithmic design approaches typically do not consider the implementation issues and constraints of the target physical platforms, in particular, the unavoidable computational and communication limitations imposed by them. This poses a problem when dealing with CPSs with complex dynamics and uncertainty. In fact, in such cases the effectiveness of designed algorithms can be compromised by the unavoidably nonzero time needed to perform computations and the quality of information transmission. The decentralization of computational resources and other effects introduced at the implementation stage that were neglected at design can negatively affect the behavior induced by the algorithm.
To properly cope with such issues, techniques for the synthesis of algorithms should incorporate information about the computation and communication requirements for their implementation, and, in some cases, possibly accept a degradation of performance while guaranteeing certain fundamental properties of the entire CPS, such as resilience, robustness, stability, and safety.
The development of such synthesis techniques requires a radical change in the way algorithms for CPSs are designed.
The main goal of this workshop is to lay out the foundations of a framework for computation and communication-aware algorithmic design of CPSs, in which, rather than being added a posteriori, computation and communication are intrinsic – in the sense that the time and cost to compute and the networked structure and communication constraints are part of the design process. We bring together to this purpose experts in CPSs and key topics on data-driven design, optimization, computer science, control, information theory, and safety.
The proposed workshop will expose attendees to cutting edge research in the field, with an eye on both theory and applications, providing students new to the field with insight on the challenges posed by computation and communication constraints on algorithm design for verification and control of CPSs and on promising directions of research.
It will also encourage collaborations between researchers on the development of algorithms for CPSs with performance guarantees.
Target Audience and Prerequisites:
The workshop aims at attracting graduate students and researchers with an interest in CPS verification and control, to get them exposed to the challenges posed by computation and communication to CPS design. The target audience also includes researchers that are more inclined towards applications than to theoretical aspects.
There are no particular prerequisites for attendees, and any graduate-level control engineer and computer scientist should be able to benefit from attending, not only grasping the main concepts but also achieving some deep level of understanding. It would be beneficial, however, if attendees have a strong mathematical background, possibly being familiar with control and systems theory. Attendees from other disciplines than control (e.g., computer science, operations research, robotics) are also qualified to attend the workshop, since the talks are expected to be self-contained, introducing the necessary prerequisites.
The workshop is organized in three sessions:
- Session 1 – Introduction and Motivation
- Session 2 – Scalability in CPS Analysis and Design
- Session 3 – Communication-constrained CPS Design
Session 1 includes an introduction to computationally-aware CPSs and a motivating application in automated transport systems.
Scalability issues in analysis and design of large scale systems are discussed in Session 2, where possible solutions are proposed for stochastic analysis and enforcement of complex specifications via symbolic control.
Session 3 deals with information-theoretic results to address the challenges arising in networked control systems due to the limited capacity of communication channels, in deterministic as well as in stochastic systems. The last presentation in Session 3 concerns the emerging Internet-of-Things (IoT) application domain where multiple devices for sensing, processing, and actuation are joined via low-cost wireless connectivity. Power allocation, uncertain channel sharing, and also security issues are considered.
The closing remarks summarize the main points and outline perspective for future research.