The Schedule

The role of domain-specific languages for cyber-physical systems

Prof. Dr.-Ing. Jeronimo CASTRILLON (Technische Universität Dresden, Germany)

Abstract: Embedded and cyber-physical systems (CPS) are heterogeneous interconnected computing systems with an ever increasing complexity. As CPSs become more widespread, the developer community widens, exposing the complexity to mainstream programmers. In this talk, I will talk about domain-specific languages (DSLs) as a promising avenue to handle complexity without compromising on efficiency. The talk will provide background on programming languages and go over sample DSLs from different communities. An in-depth example will serve to grasp the power that lies in DSLs for efficiency, correctness and ease to target complex emerging systems.  

Deterministic Actors

By Prof. Edward A. LEE (University of California at Berkeley, USA)

Abstract: The design of concurrent, real-time, and distributed software for embedded systems, robotics, and the internet of things has been evolving, moving away from low-level C code and RTOS scheduling. Increasingly promising frameworks based on publish-and-subscribe (e.g. ROS, MQTT), service-oriented architectures (e.g. gRPC, Apache Thrift), or actors (e.g. Erlang, Ray, Akka) offer higher-level abstractions with better control over concurrency. However, these technologies have been developed for or modeled after enterprise-scale information technology and have not been adapted to the unique requirements of cyber-physical systems. In particular, they have nondeterministic concurrency and weak control over timing. Moreover, these technologies are not well poised to take advantage of impending technology improvements in time-sensitive networking and precision-timed microprocessors.

In this talk, I will introduce Lingua Franca, a polyglot coordination language with an explicit model of time, more deterministic concurrency, and support for efficient, fault-tolerant, distributed applications. In Lingua Franca, components called reactors (actors revisited) execute under a deterministic, discrete-event model of computation that combines the best features of actors with the best features of synchronous languages. The functionality of a reactor is written in an unmodified target language (currently C, C++, or TypeScript). Using the C target, the Lingua Franca compiler generates extremely efficient, low footprint embedded C code that can execute on an embedded bare-iron platform or on a high-end multicore microprocessor, transparently exploiting application parallelism and realizing earliest-deadline-first scheduling. With the TypeScript target, seamless integration with the Node.js ecosystem offers a wealth of high-level IoT capabilities.

The Lingua Franca design team currently consists of Marten Lohstroh, Christian Menard, Soroush Bateni, Matt Weber, Alexander Schulz-Rosengarten, Shaokai Lin, and Edward Lee, with smaller contributions from a number of others. The language and implementation are open source with a BSD license. Many aspects of the language design are based on decades of experience with the Ptolemy II framework.

Cyber-Physical Systems: Secure-By-Design

By Prof. Anupam CHATTOPADHYAY (Nanyang Technological University, Singapore)

Abstract: With continuous growth of communication technologies and data-driven decisioning through deep learning, modern cyber physical systems are experiencing increasing autonomy and decentralization, opening up attractive application segments like smart home, smart grid, autonomous vehicles. An overlooked issue in this fast-track technology adoption is security and privacy, which we will discuss in this presentation. As we will show with some case studies, without proper design and planning in place, it is surprisingly easy to ‘attack’ small and large-scale cyber-physical systems. Consequently, an approach of secure-by-design is advocated, which we will discuss in detail in the context of (semi-)autonomous vehicle. We will finally present some early-stage simulation and prototyping platforms that aid in this secure-by-design methodology. 

Timing Considerations in Learning-Enabled Safety-Critical Systems

By Prof. Sanjoy BARUAH (Washington University in ST. Louis, USA)

Abstract: Many safety-critical cyber-physical systems are required to have the safety of their run-time behavior assured prior to their deployment. We will examine how the increasing use of Learning-Enabled Components (LECs) – software components that make use of Deep Learning and related AI technologies – in safety-critical cyber-physical systems throws up new challenges to obtaining such assurance. There are multiple aspects to a system’s safety, including functional safety (the system should perform the correct function) and timing safety (it should do so at the correct time).   We will investigate approaches for successfully assuring the safety of systems that use LECs despite these novel challenges, with a particular emphasis on the verification of timing safety properties.

Stability of feedback control in wireless systems

By Dr.-Ing. Marco ZIMMERLING (Technische Universität Dresden, Germany)

Abstract: A large number of wirelessly interconnected sensors and actuators dispersed into the environment and interacting with computing elements can help solve societal-scale problems, from sustainable cities to personalized medicine. Referred to as cyber-physical systems, they offer an unprecedented ability to monitor and act on the physical world.

In this talk, I will begin by illustrating some of the opportunities and challenges of cyber-physical systems. Using smart manufacturing as an example application, I will highlight the need to coordinate distributed and mobile sensors and actuators in a predictable and adaptive, yet highly efficient manner over low-cost, unreliable wireless networks. I will then focus on closed-loop stability, a crucial predictability property, and explain why providing formal stability guarantees can be hindered by typical wireless network imperfections, including message loss, jitter, and delays. A high-level overview of our own work will illustrate how some of these imperfections can be mitigated to the extent that well-known control techniques can be used to ultimately provide stability guarantees. I will conclude with perspectives on open problems and emerging topics