The Operational Design Domain (ODD) is a representative model of the real world in which an Automated Driving System (ADS) is intended to operate. The definition of the ODD is a crucial part of the development process for such an artificial intelligence (AI)-enabled system. This is due to the fact that the ODD is the basis for several critical development activities, like defining system-level requirements, test & verification, and building a well-founded safety case for an AI-based ADS. Since an inadequately defined ODD poses a major safety concern for the entire development, an ODD must be defined completely and consistently during the development process. In this work, we present an approach for the ODD definition and maintenance during the development of safety-critical AI-based ADS functionalities and provide evidences to argue the sufficient completeness and consistency. We demonstrate the feasibility of our approach by an industrial use case of a fully automated system in the railway domain.

Traditional automation technologies alone are not sufficient to enable the fully automated operation of trains. However, Artificial Intelligence (AI) and Machine Learning (ML) offers great potential to realize the mandatory novel functions to replace the tasks of a human train driver, such as obstacle detection on the tracks. The problem, which still remains unresolved, is to find a practical way to link AI/ML techniques with the requirements and approval processes that are applied in the railway domain. The safe.trAIn project aims to lay the foundation for the safe use of AI/ML to achieve the driverless operation of rail vehicles and thus addresses this key technological challenge hindering the adoption of unmanned rail transport. The project goals are to develop guidelines and methods for the reliable engineering and safety assurance of ML in the railway domain. Therefore, the project investigates methods to reliable design ML models and to prove the trustworthiness of AI-based functions taking robustness, uncertainty, and transparency aspects of the AI/ML model into account.

Digital transformation fundamentally changes established practices in public and private sector. Hence, it represents an opportunity to improve the value creation processes (e.g., “industry 4.0”) and to rethink how to address customers' needs such as “data-driven business models” and “Mobility-as-a-Service”. Dependable, collaborative and autonomous systems are playing a central role in this transformation process. Furthermore, the emergence of data-driven approaches combined with autonomous systems will lead to new business models and market dynamics. Innovative approaches to reorganise the value creation ecosystem, to enable distributed engineering of dependable systems and to answer urgent questions such as liability will be required. Consequently, digital transformation requires a comprehensive multi-stakeholder approach which properly balances technology, ecosystem and business innovation. Targets of this paper are (a) to introduce digital transformation and the role of / opportunities provided by autonomous systems, (b) to introduce Digital Depednability Identities (DDI) - a technology for dependability engineering of collaborative, autonomous CPS, and (c) to propose an appropriate agile approach for innovation management based on business model innovation and co-entrepreneurship.

A characteristic of reinforcement learning is the ability to develop unforeseen strategies when solving problems. While such strategies sometimes yield superior performance, they may also result in undesired or even dangerous behavior. In industrial scenarios, a system's behavior also needs to be predictable and lie within defined ranges. To enable the agents to learn (how) to align with a given specification, this paper proposes to explicitly transfer functional and non-functional requirements into shaped rewards. Experiments are carried out on the smart factory, a multi-agent environment modeling an industrial lot-size-one production facility, with up to eight agents and different multi-agent reinforcement learning algorithms. Results indicate that compliance with functional and non-functional constraints can be achieved by the proposed approach.

The MobSTr dataset contains a number of artifacts for an autonomous driver assistance system, ranging from textual requirements to models for system design and models relevant to safety assurance. The artifacts provided are connected with traceability links created and managed with Eclipse Capra, an open source traceability management tool. The dataset builds upon a custom traceability information model that provides type safety and semantics for the trace links. MobSTr is intended for researchers that work on software and systems traceability as well as on model-based safety assurance. It is already being used in a number of studies, including research on trace link consistency, change impact analysis, and automated analysis of safety and timing requirements.

Abstract The future will encompass heavily interconnected, distributed, heterogeneous and intelligent systems which are bound to have a significant economic and social impact. Cyber physical systems (CPS) such as autonomous cars, smart electric grid, implanted medical devices and smart manufacturing are some practical examples of these intelligent systems. However, due to the open and cooperative nature of CPS, assuring their dependability is a challenge. The DEIS project addresses this important and unsolved challenge by developing the concept of a digital dependability identity (DDI). A DDI contains all the information that uniquely describes the dependability characteristics of a CPS or CPS component. DDIs are synthesised at development time and are the basis for the (semi)automated integration of components into systems during development, as well as for the fully automated dynamic integration of systems into systems of systems in the field.

Abstract Systems are becoming increasingly more adaptive, using techniques like machine learning to enhance their behavior on their own rather than only through human developers programming them. We analyze the impact the advent of these new techniques has on the discipline of rigorous software engineering, especially on the issue of quality assurance. To this end, we provide a general description of the processes related to machine learning and embed them into a formal framework for the analysis of adaptivity, recognizing that to test an adaptive system a new approach to adaptive testing is necessary. We introduce scenario coevolution as a design pattern describing how system and test can work as antagonists in the process of software evolution. While the general pattern applies to large-scale processes (including human developers further augmenting the system), we show all techniques on a smaller-scale example of an agent navigating a simple smart factory. We point out new aspects in software engineering for adaptive systems that may be tackled naturally using scenario coevolution. This work is a substantially extended take on Gabor et al. (International symposium on leveraging applications of formal methods, Springer, pp 137–154, 2018).

Short product lifecycles and a high variety of products force industrial manufacturing processes to change frequently. Due to the manual approach of many quality analysis techniques, they can significantly slow down adaption processes of production systems or make production unprofitable. Therefore, automating them can be a key technology for keeping pace with market demand of the future. The methodology presented here aims at a meta-model supporting automation for PFMEA. The method differentiates product requirements, production steps and quality measures in such a way, that complex quality requirements can be addressed in any instance of a factory using a common meta-modeling language.

The growing size and complexity of software in embedded systems poses new challenges to the safety assessment of embedded control systems. In industrial practice, the control software is mostly treated as a black box during the system's safety analysis. The appropriate representation of the failure propagation of the software is a pressing need in order to increase the accuracy of safety analyses. However, it also increase the effort for creating and maintaining the safety analysis models (such as fault trees) significantly. In this work, we present a method to automatically generate Component Fault Trees from Continuous Function Charts. This method aims at generating the failure propagation model of the detailed software specification. Hence, control software can be included into safety analyses without additional manual effort required to construct the safety analysis models of the software. Moreover, safety analyses created during early system specification phases can be verified by comparing it with the automatically generated one in the detailed specification phased.

Cyber-Physical Systems (CPS) provide enormous potential for innovation but a precondition for this is that the issue of dependability has been addressed. This paper presents the concept of a Digital Dependability Identity (DDI) of a component or system as foundation for assuring the dependability of CPS. A DDI is an analyzable and potentially executable model of information about the dependability of a component or system. We argue that DDIs must fulfill a number of properties including being universally useful across supply chains, enabling off-line certification of systems where possible, and providing capabilities for in-field certification of safety of CPS. In this paper, we focus on system safety as one integral part of dependability and as a practical demonstration of the concept, we present an initial implementation of DDIs in the form of Conditional Safety Certificates (also known as ConSerts). We explain ConSerts and their practical operationalization based on an illustrative example.

In this work, we consider reliable runtime adaptation in networked, embedded systems with tight real-time constraints. Specifically, we adapt the placement of software components on a multitude of hardware components and show the need for a hierarchical transaction concept in this context. We consider multiple adaptation steps under these hard system constraints and also introduce a model with undesired configurations, which cannot be maintained for an extended time period. Furthermore, we identify cases when adaptation steps can be performed in parallel. Our concept for adaptations is applied to realistic automotive examples, where the feasibility of adaptation in the inactive period of a software component is demonstrated.

The complexity of modern vehicular embedded systems is constantly rising. In addition, distributed embedded systems like automobiles often implement safety-relevant applications which have a high demands on safety and reliability. This poses a great challenge for the design of these systems. Self-adaptation may overcome these challenges and enhance the flexibility and robustness of automotive embedded systems. To design such systems in an efficient way, an adaptive system has to be verified and validated even in early stages of the development process. Co-simulation enables such an approach. In this paper, we outline a concept for iterative virtual prototyping of the entire automotive in-vehicle network including hardware components, software functions and interconnection networks. Furthermore, we present an approach to simulate self-adaptive behavior of the automotive embedded system.