Computer Safety, Reliability, and Security. SAFECOMP 2021 Workshops: DECSoS, MAPSOD, DepDevOps, USDAI, and WAISE, York, UK, September 7, 2021, Proceedings (Lecture Notes in Computer Science)

This book constitutes the proceedings of the Workshops held in conjunction with SAFECOMP 2021, the 40th International Conference on Computer Safety, Reliability and Security, which took place in York, UK, in September 2021. The 26 regular papers included in this volume were carefully reviewed and selected from 34 submissions. The workshops included in this volume are: DECSoS 2021: 16th Workshop on Dependable Smart Embedded and Cyber-Physical Systems and Systems-of-Systems WAISE 2021 : Fourth International Workshop on Artificial Intelligence Safety Engineering DepDevOps 2021 : Second International Workshop on Dependable Development-Operation Continuum Methods for Dependable Cyber-Physical Systems USDAI 2021 : Second International Workshop on Underpinnings for Safe Distributed AI MAPSOD 2021 : First International Workshop on Multi-concern Assurance Practices in Software Design Preface Organization Contents 16th International ERCIM/EWICS/ARTEMIS Workshop on Dependable Smart Embedded Cyber-Physical Systems and Systems-of-Systems (DECSoS 2021) 16th International Workshop on Dependable Smart Cyber-Physical Systems and Systems-of-Systems (DECSoS 2021) 1 Introduction 2 ECSEL: The European Cyber-Physical Systems Initiative 3 This Year’s Workshop 4 Acknowledgements International Program Committee Dependable Integration Concepts for Human-Centric AI-Based Systems 1 Introduction 2 Related Work and Regulation Activity Overview 2.1 Research Related Work 2.2 Regulations and Standards for Automated Vehicles 2.3 Regulations and Standards for AI-Based Systems 2.4 Consequential Challenges 3 Conceptual Approaches for Ensuring Dependability of AI-Based Systems 3.1 Concept A: Human-in-the-Loop 3.2 Concept B: Policy-Based Decision Integration 3.3 Concept C: Model-Based Decision Integration 3.4 Concept D: AI-Based System for Ensuring Dependability 4 Conclusion and Outlook References Rule-Based Threat Analysis and Mitigation for the Automotive Domain 1 Introduction 1.1 Related Work 2 Structure of the Rule-Based Approach 2.1 Threat Analysis with ThreatGet 2.2 Inferring Threat Mitigation 3 Architectural Aspects of Autonomous Driving 4 Case Study: ADAS Development Platform 5 Establishing Relationships with Ontologies 6 Results 6.1 Threats and Risks Discovered 6.2 Traceability from Threats to Security Requirements 7 Conclusion References Guideline for Architectural Safety, Security and Privacy Implementations Using Design Patterns: SECREDAS Approach 1 Introduction 2 The Guideline 3 Application Example 4 Conclusion References Structured Traceability of Security and Privacy Principles for Designing Safe Automated Systems 1 Introduction 2 Problem Statement 2.1 Proposed Structure to Trace Principles to System Components 3 Use Case Example 3.1 Stakeholders and System Parameters 3.2 Principles and Security and Privacy Technology Elements 3.3 Technology Elements to Design Elements 3.4 Holistic Analysis and Identifying the Points for Improvement 4 Discussion and Conclusions References Synchronisation of an Automotive Multi-concern Development Process 1 Introduction 2 Methodology and Scope 3 Guideline for a Multi-concern Development Process 3.1 Elaboration on Concerns 3.2 Co-design 3.3 Co-verification 4 Related Work 5 Conclusions References Offline Access to a Vehicle via PKI-Based Authentication 1 Introduction 1.1 Standard Car Access Methods 1.2 Using PKI in the Automotive Domain 1.3 PKI Problems and Limitations in Car Access Management 1.4 Alternative Approaches to Car Access 1.5 The Security of the PKI System 2 The Proposed Concept 2.1 Architecture 2.2 X.509 Adaptation 2.3 Special Situations 2.4 Evaluating the Validation Time for a Certificate 2.5 Limitations of the Study 3 Conclusion References HEIFU - Hexa Exterior Intelligent Flying Unit 1 Introduction 2 Hardware 3 Software 3.1 Perception 3.2 Collision Aware Planner 3.3 Plan Handler 3.4 Command Multiplexer 3.5 Communication Handler 4 Collision Avoidance 4.1 Static Collision Avoidance 5 Online Connectivity 6 Field Tests Results 7 Conclusions and Future Work References Testing for IT Security: A Guided Search Pattern for Exploitable Vulnerability Classes 1 Introduction 2 Related Work 3 Guided Search Pattern for Arbitrary Vulnerability Classes 4 Instantiation of the Guided Search Pattern for Buffer Overflow 4.1 Introduction of the Vulnerability Class Buffer Overflow 4.2 Guided Search Structure 4.3 Evaluation 5 Instantiation of the Guided Search Pattern for Race Condition 5.1 Introduction of the Vulnerability Class Race Condition 5.2 Guided Search Structure 5.3 Evaluation 6 Conclusion and Future Work References Formal Modelling of the Impact of Cyber Attacks on Railway Safety 1 Introduction 2 Safety and Security Analysis of Moving Block 3 Event-B 4 Moving Block Development in Event-B 5 Related Work 6 Conclusion References LoRaWAN with HSM as a Security Improvement for Agriculture Applications - Evaluation 1 Introduction 2 Security Features Provided by an HSM 3 Event Signalisation: Alarms and Notifications 4 SED Evaluation of Security Measures 4.1 Detection of Unauthorised Moving of the Sensor 4.2 Ensuring the Physical Integrity of the Sensor 4.3 Inhibit Unauthorised Reuse of Manipulated Sensors 4.4 Prevent Manipulation of the Sensor Communication Data 5 Cost of Security 6 Conclusions 7 Outlook References 2nd International Workshop on Dependable Development-Operation Continuum Methods for Dependable Cyber-Physical System (DepDevOps 2021) 2nd International Workshop on Dependable Development-Operation Continuum Methods for Dependable Cyber-Physical Systems (DepDevOps 2021) 1 Introduction 2 H2020 Projects: Dependable DevOps 3 Acknowledgments International Program Committee Towards Continuous Safety Assessment in Context of DevOps 1 Introduction 2 Related Work 3 Challenges in Enabling Continuous Delivery for Safety-Critical Systems 4 Methods to Enable Continuous Delivery for Safety-Critical Systems 5 Continuous Safety Assessment 6 Towards DevOps for Safety-Critical Systems 7 Conclusions and Future Work References The Digital Twin as a Common Knowledge Base in DevOps to Support Continuous System Evolution 1 Introduction 2 Example System 3 Approach 3.1 Variant Detection 3.2 Release Management and Virtual Commissioning 4 Application of the Approach 5 Discussion 6 Related Work 7 Conclusion References 1st International Workshop on Multi-concern Assurance Practices in Software Design (MAPSOD 2021) 1st International Workshop on Multi-concern Assurance Practices in Software Design (MAPSOD 2021) 1 Introduction 2 This Year’s Workshop 3 Acknowledgements International Program Committee An Accountability Approach to Resolve Multi-stakeholder Conflicts 1 Introduction 2 Background 3 A Proposal of Accountability Map 3.1 Development of the Accountability Map 3.2 Identify Steps to Implement Accountability Using Accountability Maps 4 Verification of Accountability Map 5 Discussion 6 Conclusions References Towards Assurance-Driven Architectural Decomposition of Software Systems 1 Introduction 2 Challenges of Assurance of Monolithic Systems 2.1 Automated Logical Reasoning 2.2 Partial Correctness 2.3 Performance Assurance 3 A 4-Dimensional Meta-architecture 4 Example: A Simple File System 5 Related Work 6 Conclusion and Future Work References 2nd International Workshop on Underpinnings for Safe Distributed Artificial Intelligence (USDAI 2021) 2nd International Workshop on Underpinnings for Safe Distributed AI (USDAI 2021) 1 Introduction 2 This Year’s Workshop 3 Acknowledgements International Program Committee Integration of a RTT Prediction into a Multi-path Communication Gateway 1 Introduction 2 Related Work 3 RTT Prediction Pipeline 3.1 Data Recording 3.2 Data Preprocessing 3.3 Feature Selection 3.4 Prediction Module Training 3.5 Model Evaluation 3.6 Model Porting 3.7 Model Execution 4 Overview of the Multi-path Communication Gateway 5 Integration of the Pipeline into the Gateway 6 Summary and Outlook on Future Work References 4th International Workshop on Artificial Intelligence Safety Engineering (WAISE 2021) 4th International Workshop on Artificial Intelligence Safety Engineering (WAISE 2021) 1 Introduction 2 This Year’s Workshop 3 Acknowledgements Improving Robustness of Deep Neural Networks for Aerial Navigation by Incorporating Input Uncertainty 1 Introduction 2 Proposed Method 2.1 Autonomous Aerial Navigation with DNNs 2.2 Uncertainty Estimation with Stochastic Variable Inputs 3 Early Experiments 4 Conclusion References No Free Lunch: Overcoming Reward Gaming in AI Safety Gridworlds 1 Introduction 2 Background 2.1 AI Safety Gridworlds Environments 3 Two Heuristics: Detecting a Free Lunch 3.1 Direct Penalty: Assuming This is a Free Lunch 3.2 Doubt: Checking if this is a Free Lunch 4 Evaluation 4.1 Case 1: Observation Tile only 4.2 Case 2: Sprinkler Tile Only 4.3 Cases 3, 4, 5: Distinguishing Observation and Sprinkler Tile 5 Related Work 6 Conclusion References Effect of Label Noise on Robustness of Deep Neural Network Object Detectors 1 Introduction 2 Related Works 3 Object Detection Loss Functions 4 Experiments and Results 4.1 Noise Robustness of the Two Losses: CE vs. FL 4.2 Effect of the Gamma Parameter () 5 Conclusion References Human-in-the-Loop Learning Methods Toward Safe DL-Based Autonomous Systems: A Review 1 Introduction 2 Background 2.1 Safety Concerns in AI-Based Autonomous Systems 2.2 Human-in-the-Loop Learning Approaches 3 Review of Human-in-the-Loop Learning Approaches 3.1 Learning from Demonstration (LfD) 3.2 Learning from Intervention (LfI) 3.3 Learning from Evaluation (LfE) 3.4 Gaps and Opportunities for Future Work 4 Conclusion References An Integrated Approach to a Safety Argumentation for AI-Based Perception Functions in Automated Driving 1 Introduction 2 Integrated Approach 2.1 Overview 2.2 Specification of the Functions and Architecture (Fig. 1, Top Left) 2.3 Specification of the Operational Design Domain (Fig. 1, Top Left) 2.4 From Hazard Analysis to Safety Goals and Safety Requirements for the AI Function (Fig. 1, Top Right) 2.5 Safety Argumentation and Assurance Case (Fig. 1, Bottom Right) 2.6 DNN Safety Measures and Metrics (Fig. 1, Bottom Left and Center) 3 Summary and Outlook References Experimental Conformance Evaluation on UBER ATG Safety Case Framework with ANSI/UL 4600 1 Introduction 2 Background 2.1 Uber ATG Safety Case Framework 2.2 ANSI/UL 4600 3 Experimental Conformance Analysis 3.1 Sample Analysis Results 3.2 Summary for Entire Analysis 3.3 Structural Analyses 4 Related Work 5 Concluding Remarks References Learning from AV Safety: Hope and Humility Shape Policy and Progress 1 Measurement Muddle 1.1 Roadmanship 2 Thresholds and Processes 2.1 Thresholds 2.2 Processes 3 Policy, Technology, and Managing AV Uncertainty References Levels of Autonomy and Safety Assurance for AI-Based Clinical Decision Systems 1 Introduction 2 Levels of Autonomy of AI Applications in Healthcare 3 Difference Between Supervised and Reinforcement Learning Applications References Certification Game for the Safety Analysis of AI-Based CPS 1 Introduction 2 Definitions and Preliminaries 3 Certification Game 3.1 Manufacturer Action: Provide_Information 3.2 Certification Agent Action: Check_Learnability 3.3 Terminal State and Safety Analysis 4 Evaluation 4.1 Step 1 Repeated by Manufacturer: New Sensor 4.2 Step 2 Repeated by Certification Agent: Model Extraction and Update Safety Analysis 5 Discussion 6 Conclusions References A New Approach to Better Consensus Building and Agreement Implementation for Trustworthy AI Systems 1 Introduction 2 Our Approach 2.1 Consensus Building and Agreement Implementation with an Assurance Case 2.2 Consensus Building Process 3 CBAC Approach Experimentation 3.1 Experiment 3.2 Result 3.3 Discussion 4 Conclusions References Author Index