18th European Dependable Computing Conference
12-15 September 2022
- AI4RAILS - Artificial Intelligence for RAILwayS
- CARS - Critical Automotive Applications: Robustness & Safety
- DREAMS - Dynamic Risk managEment for Autonomous Systems
- SERENE - Software Engineering for Resilient Systems
AIR4RAILS - 3rd International Workshop on Artificial Intelligence for RAILwayS
Valeria Vittorini (University of Naples Federico II, Italy)
Francesco Flammini (Mälardalen University & Linnaeus University, Sweden)
Stefano Marrone (University of Naples Federico II, Italy)
Roberto Nardone (University of Naples "Parthenope", Italy)
The impact of Artificial Intelligence (AI) on the industry has been so disruptive that it gave rise to a new wave of research and applications. Several industries, mainly in logistics and manufacturing, have benefited significantly from AI adoption and this positive trend is planned to be kept in the future. Modern and future railways represent one of the fields in which AI is expected to have a significant impact in a medium to long term perspective, to get higher levels of automation. Despite this opening for unprecedented scenarios in railway systems, it also raises concerns regarding system dependability and new threats associated with a high level of autonomy. Therefore, the first step towards the adoption of AI in the railway sector is understanding to what extent AI can be considered reliable, safe and secure. In such a context, building upon the ongoing Shift2Rail project RAILS (Roadmaps for the AI integration in the raiL Sector), in 2019 the AI4RAILS workshop series has been started with the aim to provide annually a forum for researchers, practitioners and business leaders to discuss and share new perspectives, ideas, technologies and solutions for effective and dependable integration of AI techniques in rail-based transportation systems.
AI4RAILS 2022 is the 3rd workshop in the series, designed to gather the best idea, comments and experiences related to the adoption of AI in railways. The series is supported by an ongoing Shift2Rail project, named RAILS (Roadmaps for the AI integration in the raiL Sector), and represents an open forum for researchers working on these themes. In fact, the Advisory Board of the RAILS project will be invited since it includes railway stakeholders and companies operating in both the AI and railway market as well as the entire set of companies and research centres of the Shift2Rail community. The themes of the workshop are relevant for the EDCC community and consistent with its aims and scope. The railway domain represents a critical domain in which all the aspects of dependability and security should be continuously addressed as a consequence of technical and technological progress. In the context of AI adoption, current literature includes different attempts of adopting AI techniques for increasing the automation level and the overall dependability of railways. AI4RAILS represents a unique forum for researchers, practitioners and business leaders to present and discuss their latest research results on theory, approaches and tools for the design, validation, operation and evaluation of dependable and secure railways by means of AI techniques.
Topics of interest include, but are not limited to:
- AI for railway safety, availability, reliability and security
- Transferability to railways of AI approaches from other sectors (e.g., avionics, automotive, etc.)
- Autonomous and cooperative driving, including virtual coupling
- Automatic detection of signals, pedestrians and obstacles
- Predictive maintenance and defect detection
- AI for traffic planning and management
- Adversarial perturbations to AI with a focus on railway case-studies
- Resource allocation with a focus on railway case-studies
- Run-time model checking and AI with a focus on railway case-studies
- Explainable and trustworthy AI in railway applications
- Safe autonomy standards and railway certification of compliance
- Datasets and real-world case studies
CARS - 7th International Workshop on Critical Automotive Applications: Robustness & Safety
Rolf Johansson (Astus AB, Sweden)
Jean-Charles Fabre (LAAS-CNRS, France)
Mario Trapp (Fraunhofer IKS, Germany)
The increasing complexity of automotive applications, the challenges posed by autonomous vehicles, the need to master production costs using off-the-shelf components, the coexistence of critical and non-critical applications, and the emergence of new architectural paradigms may have a strong effect on dependability of automotive embedded systems. This situation requires design and validation methods, but also tools to improve automotive systems robustness and their safety and security properties.
The evolution of automotive systems is supported by standards (like AUTOSAR, ISO26262) that respectively advocate a software architecture but also recommend development approaches to improve system robustness and safety. The fast evolution of both standards is a reality to incorporate novel features, more flexibility while improving robustness, security and safety.
The goal of the workshop is to bring together researchers and practitioners interested in the construction of critical automotive applications and systems. It will place emphasis on:
- Dependability issues
- Software engineering for robustness
- Security and safety issues
- Real-time embedded systems technologies
- Architectural software and hardware solutions
- Development processes for dependable automotive embedded systems
In particular, CARS aims at promoting and fostering discussion on novel ideas and techniques, possibly controversial approaches, it is a forum for on-going work exchange, a place where researchers and developers can share both real problems and innovative solutions. Application areas of interest to the workshop focus on the automotive domain but methods and techniques in other transport domains (e.g. aerospace, railways) are also welcome. The workshop is complementary to the EDCC main track as it focuses on dependability in automotive embedded systems. The intended duration for the workshop is one full day.
DREAMS - 3rd International Workshop on Dynamic Risk managEment for Autonomous Systems
Rasmus Adler, Philipp Schleiß (Fraunhofer IESE, Germany)
Richard Hawkins (University of York / Assuring Autonomy International Program, UK)
Autonomous systems have enormous potential and they are bound to be a major driver in future economical and societal transformations. In contexts where safety, or other critical properties, need to be guaranteed it is, however, presently not possible to exploit autonomous systems to their full potential. Unknowns and uncertainties are induced due to the high complexity of the autonomous behaviors, the utilized technology and the volatile and highly complex system contexts. These characteristics render the base assumptions of established assurance methodologies (and standards) void, hence new approaches need to be investigated. One general approach for making autonomous systems dependable is to make them aware of risks and empower them to assess and control those risks. Implementing such a Dynamic Risk Management (DRM) approach comes along with many challenges concerning the necessary self- and context awareness. On the one hand, powerful and thus complex self- and context awareness is necessary to minimize risks, to resolve conflicting objectives and to make acceptable trade-off decisions. On the other hand, the complexity of DRM is in conflict with assurance and high confidence in adequate risk management. DRM has the potential to not only outright enable certain types of systems or applications, but also to significantly increase the performance of already existing ones. This is due to the fact that by resolving unknowns and dealing with uncertainties at runtime it will be possible to get rid of worst-case assumptions that are typically detrimental to the systems performance properties.
The DREAMS workshop intends to explore concepts, techniques and technology for realizing DRM. It invites experts, researchers, and practitioners for presentations and in-depth discussions about the current status of DRM in practice, its relevance for specific use cases, its relation to exiting assurance frameworks for autonomous systems and standardization activities. DREAMS aims at bringing together communities from diverse disciplines, such as safety engineering, runtime adaptation, predictive modelling, control theory, and from different application domains such as automotive, healthcare, manufacturing, agriculture and critical infrastructures.
Topics of interest include but are not limited to:
- Definition and modelling of the Operational Design Domain
- Approaches for monitoring AI-based components
- (Runtime) estimation and handling of uncertainties
- Approaches to formalize risk and risk assessment
- Ethics of risk / machine ethics
- (Dynamic) assurance cases for autonomous systems and AI
- Collecting data from operation for continuous risk management / MLOps
- Case studies
SERENE - 14th International Workshop on Software Engineering for Resilient Systems
Martina De Sanctis (Gran Sasso Science Institute, Italy)
Imre Kocsis (Budapest University of Technology and Economics, Hungary)
Resilient systems withstand, recover from, and adapt to disruptive changes with acceptable degradation in their provided services. Resilience is particularly important in modern software and software-controlled systems, many of which are required to continually adapt their architecture and parameters in response to evolving requirements, customer feedback, new business needs and platform upgrades. Resilience also has to cover extra-functional behavior: despite frequent changes and disruptions, including unforeseen failures and malicious cyber attacks, systems are expected to function correctly and reliably. This is particularly important for critical services, e.g., in transportation, healthcare, energy production and e-government. Design for resilience is an increasingly important area of software engineering; new deployment platforms such as edge and fog computing and blockchains/distributed ledgers facilitate new resilience techniques, but also pose new challenges.