Guided safety analysis for cyber physical systems
Inventors
Huang, Jane • Vierhauser, Michael • Agrawal, Ankit
Assignees
Publication Number
US-11593097-B2
Publication Date
2023-02-28
Expiration Date
2039-10-15
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
Systems and methods for maintaining the safety of a software-based system. One method includes automatically generating a first artifact tree for a hazard for a first version of the system and automatically transforming the first artifact tree into a first augmented tree using a set of heuristics. The method also includes automatically generating a second artifact tree for the hazard for a second version of the system and automatically transforming the second artifact tree for the hazard into a second augmented tree using the set of heuristics. The method further includes automatically comparing the first augmented tree and the second augmented tree to generate a delta view, and automatically generating, based on the delta view, at least one selected from a group consisting of a safety warning for the second version of the software-based system and an actionable recommendation to maintain safety of the second version of the software-based system.
Core Innovation
The invention addresses the challenge of maintaining and evolving safety cases for software-based systems, particularly those with safety-critical implications such as medical devices, UAVs, autonomous vehicles, and robots. Current tools for managing traceability and safety arguments across software versions are rudimentary, requiring manual work and making it costly and difficult to certify system changes, thus limiting innovation. The patent proposes an automated approach that helps analysts and developers build, update, and maintain structured safety arguments as the system evolves.
The core innovation lies in Safety Artifact Forest Analysis (SAFA), a process that automatically generates artifact trees from project repositories for a specified hazard in both the existing and updated versions of the software. These trees are then transformed using heuristics to add argumentation structures, resulting in augmented safety trees. By comparing safety trees of different versions, the system automatically visualizes changes as a delta view, highlights elements potentially impacting safety, and generates actionable warnings and recommendations for safety assurance case updates.
The process leverages traceability information models to organize requirements, design definitions, tests, and context-related artifacts into hierarchical structures linked via traceability paths. Automated analysis mines source repositories and requirements systems, detects changes affecting hazards, and produces recommendations for remediation, enabling evidence-based and systematic updates to safety assurance cases. This methodology significantly reduces the effort required for recertification and supports continued evolution of safety-critical software systems.
Claims Coverage
The patent claims encompass three independent inventions: a method, a system, and a non-transitory computer readable medium for maintaining the safety of a software-based system. The main inventive features are as follows.
Automated generation and transformation of artifact trees using heuristics
- Automatically generates a first artifact tree for a hazard for a first software version by recursively following traceability paths for artifacts mitigating the hazard. - Automatically transforms the artifact tree into an augmented tree using a set of heuristics that insert augmentation nodes.
Delta view comparison for safety change detection
- Automatically generates a second artifact tree for a hazard for a second version and transforms it into a second augmented tree with heuristics. - Compares the first and second augmented trees to generate a delta view. - The delta view identifies added, modified, or deleted artifacts and links between software versions.
Automated creation of candidate trace links and actionable modifications
- Automatically creates at least one candidate trace link for the software-based system based on the delta view to maintain safety of the new version. - Modifies the software-based system based on the candidate trace link. - Includes generating warnings and actionable recommendations based on the analysis.
System and non-transitory medium implementations
- Provides the same automated functions via a system with at least one electronic processor implementing the method steps. - Claims a non-transitory computer readable medium with instructions for carrying out the automated generation, comparison, and modification process.
In summary, the claims cover automated generation and augmentation of safety artifact trees, automated delta analysis for change detection, automated generation of candidate trace links and recommendations, and implementation via systems and computer-readable media.
Stated Advantages
Reduces the cost and effort required to evolve and reuse safety case elements for new versions of a system, while maintaining or increasing safety levels.
Enables identification and visualization of changes affecting system safety through automated generation of delta views, providing faster and more accurate detection than manual methods.
Provides actionable recommendations and warnings that guide analysts in evaluating and updating safety cases in response to changes.
Facilitates innovation and release of new system features by lowering the overhead of recertification and maintenance of safety assurance cases.
Supports both safety-critical and other domains by providing traceability and safety analysis tools compatible with diverse project environments.
Improves the usability and maintainability of safety cases compared to traditional trace matrices and artifact management systems.
Documented Applications
Supporting safety analysis and certification for safety-critical software systems, including medical devices, unmanned aerial vehicle (UAV) applications, autonomous vehicles, robots, signaling systems, and railway electronic systems.
Providing tools for traceability and safety case analysis in organizations building devices with safety impact, even when formal methods are not used.
Integrating with requirements management and version control tools (such as Jira, DOORS, GitHub, Bitbucket) to facilitate agile development and safety case evolution in cyber-physical systems.
Application demonstrated in monitoring and coordinating unmanned aerial vehicles (UAVs) in search-and-rescue and medical supply delivery scenarios.
Supporting the maintenance and certification of safety cases in accordance with guidelines and standards such as those from the US FDA and Ministry of Defense of the UK.
Interested in licensing this patent?