Computer Safety, Reliability, and Security: 41st International Conference, SAFECOMP 2022, Munich, Germany, September 6–9, 2022, Proceedings (Lecture Notes in Computer Science)

Preface
Organization
Contents
Safety Analysis and Certification
Analysing the Safety of Decision-Making in Autonomous Systems
	1 Introduction
	2 Background
	3 A Decision Safety Analysis Process
		3.1 The Decision Safety Analysis Process
		3.2 Robot Delivery System Example
	4 Process Evaluation
	5 Discussion and Conclusions
	References
BayesianSafety - An Open-Source Package for Causality-Guided, Multi-model Safety Analysis
	1 Introduction
	2 Preliminaries
		2.1 Bayesian Networks
		2.2 Model-to-Model Transformations
		2.3 Bridging the Issue of Multiple Domains
	3 Package BayesianSafety
		3.1 Models and Their Combinations
		3.2 Model Inference
		3.3 Technical Ecosystem
	4 Related Work
	5 Conclusion
	References
Safety Certification with the Open Source Microkernel-Based Operating System L4Re
	1 Introduction
	2 Background Material
	3 Certification Approach
		3.1 Architecture of the EB Corbos Hypervisor
		3.2 General Thoughts on the Qualification Strategy
		3.3 New Safety Architecture
	4 Conclusion
	References
Data-Driven Inference of Fault Tree Models Exploiting Symmetry and Modularization
	1 Introduction
	2 Fault Trees
	3 Modules and Symmetries
		3.1 Modules
		3.2 Symmetries
	4 Exploiting Modules and Symmetries in FT Inference
	5 Experimental Evaluation
	6 Conclusions
	References
Assurance Cases
ARACHNE: Automated Validation of Assurance Cases with Stochastic Contract Networks
	1 Introduction
	2 Background
	3 Assurance Cases as Contract Networks
		3.1 Stochastic Propositional Logic
		3.2 Hierarchical Stochastic Contract Networks
		3.3 Confidence Networks
	4 Assurance Case Validation
		4.1 Checking the Soundness of an HSCN
		4.2 Confidence Assessment
	5 Case Studies
		5.1 Software Correctness Assurance
		5.2 Scalable Assurance for Autonomous Driving Systems
	6 Conclusions
	References
Automating Pattern Selection for Assurance Case Development for Cyber-Physical Systems
	1 Introduction
	2 Assurance Case Patterns
		2.1 Pattern Formalization
	3 Pattern Selection Workflow
		3.1 Data Preparation
		3.2 Pattern Selection
		3.3 Coverage Evaluation
	4 Illustrative Example
	5 Related Work
	6 Conclusion and Future Work
	References
Generating Assurance Cases Using Workflow+ Models
	1 Introduction
	2 An Example in GSN
	3 Generating GSN-Like Arguments from WF+
		3.1 Syntactic Constraints
		3.2 Semantic Constraints
		3.3 Deriving Higher-Level Argumentation
		3.4 Integrating Assurance Segments
		3.5 Instantiating WF+ Models
	4 Related Work
	5 Evaluation
		5.1 Collaboration with an Industrial Partner
		5.2 Comparison with Other Approaches
	6 Conclusion
	References
Uncertainty Elicitation and Propagation in GSN Models of Assurance Cases
	1 Introduction
	2 Background and Related Work
	3 From GSN to Dempster-Shafer Theory
		3.1 Logical Modeling of GSN
		3.2 Uncertainty Propagation Model
		3.3 Belief and Disbelief Elicitation
	4 Uncertainty Assessment Procedure
	5 Case Study
	6 Conclusion
	References
Fault Detection, Monitoring and Tolerance
Impact of Machine Learning on Safety Monitors
	1 Introduction
	2 Related Work
	3 Problem Statement
	4 System Model and Terminology
		4.1 Terminology
		4.2 Description of the State Space
	5 Study Method
		5.1 Paired Tests with and Without Safety Monitor
		5.2 Evaluation of the Components and the System (Vehicle) Safety
	6 Details of the Simulation
		6.1 CARLA Simulator
		6.2 Implementation of the Controller and Safety Monitor
		6.3 Structure of the Study
	7 Results of the Simulation
		7.1 Controller
		7.2 Safety Monitor
		7.3 Whole-Vehicle Evaluation
	8 Concluding Remarks
	References
Comprehensive Analysis of Software-Based Fault Tolerance with Arithmetic Coding for Performant Encoding of Integer Calculations
	1 Introduction
		1.1 Fault Tolerance
		1.2 Arithmetic Coding
		1.3 Contribution
	2 Encoding Strategy
	3 Comparison of Arithmetic Codes and Their Implementations
		3.1 AN Codes
		3.2 Residue-Codes
		3.3 Complement-Codes
		3.4 Comparison
	4 Fault Detection Capabilities
	5 Performance Overhead
	6 Evaluation and Experimental Validation of the Ones\' Complement
		6.1 Evaluation
		6.2 Experimental Validation
		6.3 Outlook for Future Validation
	7 Related Works
	8 Conclusion
	References
STPA-Driven Multilevel Runtime Monitoring for In-Time Hazard Detection
	1 Introduction
	2 STPA-Driven Runtime Monitor Design
		2.1 Losses, Hazards and Unsafe Control Actions
		2.2 Causal Factors and Relation to Multilevel Monitoring
		2.3 Multilevel Runtime Monitoring Framework
	3 Monitoring an AEB Controller
		3.1 STPA for AEB
		3.2 Loss Scenarios and Causal Factors as Design Guides for Multilevel Runtime Monitoring
	4 Conclusion
	References
Security and Safety
Proposal of Cybersecurity and Safety Co-engineering Approaches on Cyber-Physical Systems
	1 Introduction
	2 State of the Art
		2.1 Standards
		2.2 Co-engineering Methods
		2.3 Positioning
	3 EBIOS RM Method
		3.1 Introduction
		3.2 Five Workshops of EBIOS RM
	4 Cybersecurity/Safety Co-engineering: Two Approaches
		4.1 First Approach: Safety Contributions to Cybersecurity EBIOS RM Workshops
		4.2 Second Approach: Contributions of Cybersecurity to the Safety Demonstration
	5 Conclusion
	References
On the Feasibility and Performance of Secure OPC UA Communication with IIoT Devices
	1 Introduction
	2 Background
	3 Related Works
	4 Evaluation Platform
		4.1 Hardware and Measurement Setup
		4.2 Software Implementation
	5 Evaluation
		5.1 Key and Certificate Generation
		5.2 Connect
		5.3 Read Operations
		5.4 Subscriptions
		5.5 Memory Footprint
		5.6 Power Consumption
	6 Implications and Discussion
	7 Conclusion
	References
Fault Injection
SailFAIL: Model-Derived Simulation-Assisted ISA-Level Fault-Injection Platforms
	1 Introduction
	2 Background
		2.1 Systematic Fault Injection
		2.2 Sail: ISA Modeling Language
		2.3 FAIL*: Fault Injection Leveraged
	3 The SailFAIL Approach
		3.1 Connecting Sail and FAIL*
		3.2 Systematic Register Access Tracing
		3.3 Bit-Wise Def-Use Pruning
		3.4 Virtual Fault Spaces
	4 Evaluation
		4.1 Simulation Overheads
		4.2 Register Trace Coverage
		4.3 Efficiency Improvements by Bit-Wise Pruning
		4.4 Case Study: SDC Counts for Bubblesort
	5 Conclusion
	References
Quality of Fault Injection Strategies on Hardware Accelerator
	1 Introduction
		1.1 Context
		1.2 Contributions
	2 Abstract Semantics of Hardware Architecture
		2.1 Semantics of Atomic Components
		2.2 Semantics of Components
		2.3 Application of the Semantics to a Streaming Architecture
	3 Methodology
		3.1 Fault Model
		3.2 Identification of Failures Scenarios
		3.3 Coverage/Fault Collapsing
	4 Experiments
	5 Related Work
	6 Conclusion
	References
Assessment of the Impact of U-space Faulty Conditions on Drones Conflict Rate
	1 Introduction
		1.1 Conflict Management by the U-space
		1.2 Safety Assessment in the U-space
	2 Related Work
		2.1 Safety Assessment of UAVs
		2.2 Fault Injection for Safety Assessment
	3 Approach and Experimental Setup
		3.1 Scenarios and Missions
		3.2 Fault/Failure Model
		3.3 Safety Assessment Metrics
		3.4 Experimental Framework
	4 Results and Analysis of the Results
		4.1 Assessment of the Impact on the Conflicts
		4.2 Assessment of the Impact on the Surveillance Performance
		4.3 Discussion
	5 Threats to Validity
	6 Conclusion
	References
ACTOR: Accelerating Fault Injection Campaigns Using Timeout Detection Based on Autocorrelation
	1 Introduction
	2 Problem Analysis
		2.1 Fault-Injection Model
		2.2 Timeout Detectors
	3 Timeout Detection Using Autocorrelation
		3.1 Adaption as Timeout Detector
		3.2 FAIL* Integration
	4 Evaluation
	5 Discussion
	6 Related Work
	7 Conclusion
	References
Object Detection and Perception
Formally Compensating Performance Limitations for Imprecise 2D Object Detection
	1 Introduction
	2 Related Work
	3 Data-Driven Safe Post-processing in Addressing 2D Object Detection Imprecision
	4 Mathematically Associating the IoU Metric and Safety
		4.1 The Mathematical Connection Between IoU and Safety
	5 Connecting Motion Planners with Safety Post-processing
	6 Evaluation
	7 Concluding Remarks
	References
Architectural Patterns for Handling Runtime Uncertainty of Data-Driven Models in Safety-Critical Perception
	1 Introduction
	2 Related Work
	3 Example Use Case
	4 Architectural Patterns for Dealing with Uncertainty
	5 Simulation-Based Evaluation Approach
	6 Study Results and Discussion
	7 Summary and Conclusion
	References
Hardware Faults that Matter: Understanding and Estimating the Safety Impact of Hardware Faults on Object Detection DNNs
	1 Introduction
	2 Related Work
	3 Preliminaries
		3.1 Hardware Faults Vocabulary
		3.2 Experimental Setup: Models, Datasets and System
	4 Methodology of Vulnerability Estimation
		4.1 Issues with Average Precision
		4.2 Proposed Metrics: IVMOD
	5 Transient Faults
		5.1 Corruption Probability
		5.2 Corruption Severity
		5.3 Bit-Wise Analysis of False Object Count
	6 Permanent Faults
		6.1 Evaluating Fault Persistence
		6.2 Corruption Probability and Severity
	7 Conclusion
	References
Application of STPA for the Elicitation of Safety Requirements for a Machine Learning-Based Perception Component in Automotive
	1 Introduction
	2 Related Work
	3 Safety Requirement Elicitation for Pedestrian Detection Component
		3.1 Integrating STPA into the Safety Life Cycle of ISO 26262 and ISO/FDIS 21448
		3.2 Case Study: Deriving Safety Requirements of a Perception Component with STPA
		3.3 Defining Safety-Relevant Metrics
		3.4 Discussion
	4 Conclusion
	References
Testing
Exploring a Maximal Number of Relevant Obstacles for Testing UAVs
	1 Introduction
	2 Scenario-Based Testing of UAVs
	3 Automated Derivation of a Maximal Number of Relevant Obstacles
		3.1 Relevant Obstacles for the UAV
		3.2 Parameter Values Describing the UAV\'s Behavior
		3.3 Process Overview of the Automated Derivation of a Maximal Number of Relevant Obstacles
		3.4 Search Space and Fitness Function
	4 Experiments
		4.1 Experimental Setup and Implementation
		4.2 Experimental Results
		4.3 Discussion
		4.4 Threats to Validity
	5 Related Work
	6 Conclusion
	References
Data-Driven Assessment of Parameterized Scenarios for Autonomous Vehicles
	1 Introduction
	2 Scenario-Based Testing
	3 Parameterized Scenario Assessment
		3.1 Extracting Relevant Information from Traffic Data
		3.2 Containment of a Single Recorded Scenario Instance
		3.3 Containment of Many Recorded Scenario Instances
	4 Experiments
		4.1 Parameterized Scenario
		4.2 Traffic Data
		4.3 Points to Be Reproduced
		4.4 Fitness Function for Re-creation
		4.5 Simulation Setup
		4.6 Experiment Results
		4.7 Discussion
		4.8 Limitations
	5 Related Work
	6 Conclusion
	References
Optimising the Reliability that Can Be Claimed for a Software-Based System Based on Failure-Free Tests of Its Components
	1 Introduction
	2 Confidence Bounds from Component Tests
	3 Optimising Test Plans for Asymmetric Structures
	4 Test Plan Optimisation Using Linear Programming
	5 Example
	6 Discussion and Conclusions
	A Test Plan Optimisation R Script
	References
Author Index