CN / EN

搜索
Search
NEWS CENTER

NEWS CENTER

新闻中心
Check category

Accelerating Functional Safety AI Intelligence: Launch of HIRAIN FuSa AI Agent

发布时间:2024-10-18 11:34

The academic paper on HIRAIN FuSa AI Agent has been accepted by the Industrial Applications Track of the prestigious EMNLP conference in the field of Natural Language Processing, which undoubtedly recognizes the innovative capabilities of HIRAIN Technologies.

 

With the increasing complexity of automotive electronic and electrical (E/E) systems and the rapid development of autonomous driving technology, functional safety research and development (R&D) faces increasingly severe challenges, leading to rising costs. In response to this challenge, HIRAIN Technologies, leveraging its deep expertise in functional safety and cutting-edge AI technology, has successfully developed HIRAIN FuSa AI Agent, a functional safety agent capable of automatically conducting hazard analysis and risk assessment on functional safety analysis objects, formulating safety goals, deriving safety requirements through safety analysis, and continuously conducting R&D, testing, and validation activities to ensure vehicle safety.

 

HIRAIN FuSa AI Agent adopts a multi-agent system design, with each agent playing a specific role, such as the Functional Safety Manager (FuSa Manager), Functional Safety Expert (FuSa Expert), and Verification and Validation Engineer (V&V Engineer), simulating the division of labor and collaboration in real-world functional safety. The system can automate complex functional safety tasks, achieving cost-effective and efficient functional safety R&D and testing implementation.

 

Accelerating Functional Safety AI Intelligence: Launch of HIRAIN FuSa AI Agent

HIRAIN FuSa AI Agent Framework

 

▎Core Functions
•    Functional Safety Goal Setting: Based on Hazard Analysis and Risk Assessment (HARA) and ASIL classification, clearly define the safety goals that the entire vehicle must meet.
•    Functional Safety Requirement Development: Analyze and output specific functional safety requirements, ASIL levels, and safety states using safety analysis methods such as FTA, based on functional safety goals.
•    Functional Safety Testing and Validation: Develop testing strategies based on safety requirements, select appropriate testing methods and test case derivation approaches, and write test cases to ensure comprehensive test coverage.

 

▎Technical Highlights
•    Retrieval-Augmented Generation (RAG): Build a domain knowledge base and combine it with an agent reflection mechanism to efficiently handle complex and knowledge-intensive tasks.
•    Multi-Agent Collaboration: Simulate real-world divisions of labor to enhance task execution professionalism.
•    Professional Prompt Optimization: Prompts are optimized by senior functional safety engineers to ensure outputs align with business needs.
•    Continuous Learning Mechanism: The system can continuously learn and update its knowledge base to stay synchronized with cutting-edge technologies and standards.

 

▎Product Advantages
•    Efficiency Enhancement: Agents provide a high-value starting point for risk assessment, significantly improving assessment efficiency.
•    Quality Improvement: Agents provide detailed reasoning processes and examples, enhancing the quality of HARA analysis.
•    Human-AI Collaboration: Human experts and AI collaborate and divide tasks, ensuring the final assessment complies with ISO 26262 standards and is of high quality.
•    Tool API Integration: Compatible with Medini Analyze and Vector Informatik, accelerating functional safety development.
•    Real-Time Standard Updates: Automatically integrates ISO 26262 and VDA 702 standards, providing real-time compliance updates.

 

▎Quick Start
•    System Login: Click the dedicated link to quickly start using the system.
•    Functional Modules: Intuitive editing on the FuSa Manager interface with simple operations.
•    Command Generation: Upload functional requirement documents (supporting txt, Excel, PDF, and other formats) for quick response and command generation.
•    Requirement Generation: View generated results with one click and export them as Excel files.

 

Example of Generated Results

 

Welcome to leave your "Name + Company + Contact Information" in the public number backend to immediately start your HIRAIN FuSa Agent trial! The product supports subscription and private deployment.

 

References:

[1] Lu Shi, Bin Qi, Jiarui Luo, Yang Zhang, Zhanzhao Liang, Zhaowei Gao, Wenke Deng, Lin Sun. Aegis: An Advanced LLM-Based Multi-Agent for Intelligent Functional Safety Engineering. EMNLP 2024.

[2] Hopfield, J. J. (1982). Neural networks and physical systems with emergent collective computational abilities. Proceedings of the National Academy of Sciences, 79(8), 2554-2558.
[3] Hinton, G. E., & Salakhutdinov, R. R. (2006). Reducing the dimensionality of data with neural networks. Science, 313(5786), 504-507.

Pop Message