|When||20 – 22 January, 2020|
|Organizers||Francesca Palumbo (UNISS), Evgeny Shindin (IBM Research Haifa)|
|Benefits & Target||Cerbero participation to the event aims at divulgation of the CERBERO toolchain capabilities among companies and researches attending at HiPEAC 2020|
|Description||Cyber-physical systems (CPS) represent the cutting edge of the ICT. They provide control, interaction and communication over, with and among the physical and the cyber world. CPS combine computing devices, interfaces, and networks with the environment to sense and provide feedback, increasingly often implementing autonomous and intelligent interactions. Dealing with CPS development requires spanning over a plethora of heterogeneous topics, which makes the process of their definition, analysis and design a challenging task. Moreover, to interact with the environment or humans the system has to be capable of providing a certain degree of flexibility, which certainly adds to the challenges. Tools to be used span from verification to modeling, and from mapping to high-level synthesis across many levels of abstraction, and must allow a combination of physical or continuous dynamics. Last but not least, CPS should react to and handle a large variety of possible perturbations, such as change in requirements, time delays, and system failures. All these issues have been fronted in CERBERO – Cross-layer modEl-based fRamework for multi-oBjective dEsign of Reconfigurable systems in unceRtain hybRid envirOnments H2020 project (www.cerbero-h2020.eu). This tutorial is meant to represent the final event of the project, and is organized in 4 slots as follows:|
SESSION 1: The CERBERO Project – Challenges and Solutions
CERBERO overview, Francesca Palumbo (UNISS)
- Advancement on (dataflow) MoCs, Eduardo Juarez (UPM)
- CERBERO Intermediate Format CIF, Evgeny Shindin (IBM Research Haifa)
- Key Performance Indicators, Francesco regazzoni (ALaRI – USI)
- CEBERO Adaptation Loop, Eduardo de la Torre (UPM)
- Formal methods in the CERBERO Toolchain, Luca Pulina (UNISS)
SESSION 2: HANDS on “Tools Interoperability – Intra-tool interoperability using the CERBERO Interoperability Framework”
Speakers: Julio Oliveira (TN0) and Michiel van den Baar (TNO)
SESSION 3: HANDS on “Self-Adaptation over Heterogeneous Embedded Computing Infrastructures”
Speakers: Claudio Rubattu (UNISS, INSA), Carlo Sau (UNICA), Daniel Madroñal (UPM), Leonardo Suriano (UPM), Alfonso Rodriguez (UPM), Tiziana Fanni (UNISS)
SESSION 4: The CERBERO Project – Impact, Demo and Clustering
Clustering: DEIS, BONSEYES, CPSWARM
Walking UC DEMO tour
Self-healing system for Planetary Exploration, Speaker: Pablo Sánchez de Rojas(TASE)
Smart Traveling for Electric Vehicles, Speakers: Joost Adriaanse (TNO), Antonella Toffetti (CRF)
Ocean Monitoring, Speakers: Hans Myrhaug (AS), Leszek Kaliciak (AS)
1. Self-healing system for Planetary Exploration demonstrator: The demonstrator highlights the main challenges of the use case: fault detection and recovery, (self-)adaption and power monitoring and optimization at a computing platform level. The application scenario tackles the generation and validation of collision-free trajectories for a robotic arm. An optimization-based motion planning algorithm integrates with the CERBERO tools and technologies in different ways: ARTICo3 (UPM-CEI) provides dynamic partial reconfiguration, transparent scalability and adjustable redundancy. PAPIFY (UPM-CITSEM) features automatic monitoring of the hardware accelerators. MDC (UNISS and UNICA) enables virtual reconfiguration intended to provide fast-switching between different implementations of the algorithm.
2. Smart Traveling for Electric Vehicles demonstrator: During the project a demonstrator was built on top of the driving simulator of CRF (Fiat/Chrysler Research) in Turin, Italy. The demonstrator added electric vehicle simulation using battery and electric motor models and a driving assistant. The assistant monitors the car, the environment and the driver during the trip and proactively provides advice on best itineraries to drive, taking into consideration the preferences of the driver. The DynAA tool from TNO supports system-in-the-loop simulation and itineraries predictions. The MECA tool from S&T provides the planning engine of the assistant. In the demonstrator the DynAA and MECA tools are integrated in the SCANeR simulation software used by CRF. The demonstrator will allow CRF to execute test scenarios for electric vehicles were driver is supported with a driving assistant to address issues like state of charge of the battery and finding the best charging options while driving. During the demo video recording will be shown of the actual demonstrator implemented at CRF in Turin.
3. Ocean Monitoring Demonstrator: Ocean monitoring (OM) can encompass a range of purposes – from observing and tracking marine ecology and climate, to subsea maintenance of equipment. The CERBERO Ocean Monitoring use case comprises smart video-sensing (an adaptive camera) uncrewed vehicles with immersive environmental monitoring capabilities. OM adopts the CERBERO new development methodology for the complete CPS design cycle – to reduce cost and development time, improve reuse, track and improve a range of KPIs, and enable incremental prototyping. OM uses the CERBERO toolchain for fast, incremental prototyping, reduced development risk, and reduced cost. This use case applies the CERBERO adaptivity framework as its architectural pattern.