Model-Based Design for Cyber-Physical Systems and the Internet Of Things
- 2 ECTS
The focus of the class is to consider methods and tools for the co-modeling and co-design of, essentially:
- distributed/concurrent embedded applications
- parallel/multicore architectures with new interconnects (Time-sensitive Networking for IoT)
- extra-functional constraints linking the applicative requirements to the architectural guarantees (performance, safety, consumption...)
Techniques for co-simulation, design space exploration and optimization of the mapping of applications onto architectures will be studied, with a clear focus on CPS control applications interacting constantly with a physical environment (Internet of Things, Automated Driving...)
There will be two parts in the class:
- I will focus on Formal mathematical Models that underly many of the academic/industrial existing tool environments: hierarchical state and activity diagrams, structural block diagrams, and the relevant semantic annotations meant to represent extra-functional properties (Worst Case Execution Time, Funcional and Temporal Fault Trees...). I will review the main relevant analysis techniques developed over the years (classical Real-Time scheduling, Model-Checking, Failure mode analysis...)
- I will present a broad collection of modeling tools and environments for Model-Based System Engineering used in this domain by prominent industrial companies, then ask students to each explore one such tool/method/environment in more details, with oral presentation of their findings to fellow students.
(non exhaustive list: SysML/MARTE, AADL, Capella/ARCADIA, Amalthea/AutoSar, Synopsys System Architect, Ansys Twin Builder...)
The class shall reflect the need for trained engineeers in many large companies, such as Thales, Airbus, Safran, Renault, PSA, and even more numerous tool vendors in the area of Embedded Cyber-Physical Systems and Internet of Things.