The FuSaGPT chatbot was developed to improve access to the ISO 26262 standard for functional safety and to facilitate the implementation of the standard. The developed chatbot offers a user-friendly platform that provides precise answers to questions about ISO26262 and its definitions. With the help of modern language models and techniques such as Retrieval Augmented Generation (RAG), FuSaGPT facilitates working with the standard and supports the development of safety-critical systems.

The challenge

The ISO 26262 standard is extensive and technically complex, making it difficult for users to quickly find the information they need. The biggest challenges include:

• Complexity of the content: The standard contains detailed and extensive definitions that must be applied specifically in the context of functional safety.
• Access to specific information: Users often need to link relevant content from different parts of the standard in order to implement the work packages described in the standard.
• Technical expertise: Not all users have in-depth knowledge to be able to interpret the standards without problems.

FuSaGPT offers an innovative solution for flexible working with the ISO26262 standard. The main components include:

1. PDF parsing and data preparation:

• Content from the ISO 26262 PDF documents is extracted and divided into smaller, manageable text sections.
• Precise pre-processing removes irrelevant content and improves data quality.

2.Creation of text embeddings:

• The edited texts are converted into numerical representations (embeddings) that represent the context and meaning of the content.
• The use of state-of-the-art pre-trained embedding models ensures high embedding accuracy.

3.Storage in a vector database:

• The generated embeddings are stored in a vector database to enable fast and context-sensitive information retrieval.

4. Response generation with RAG technology:

• Retrieval Augmented Generation (RAG) combines the flexibility of dynamic data bases with the generalization capabilities of modern language models. Contextually relevant content is retrieved and used to generate precise answers.
• The basis is the powerful, locally executable Llama3 model, which generates high-quality, context-related answers from the relevant text sections of the standard.

5. User interface:

• With the help of Streamlit, an intuitive and lightweight user interface was developed that enables easy interaction with the chatbot.

6. Optimized prompt generation:

• Customized prompts control the chatbot's responses and improve its adaptability to different user questions.

FuSaGPT provides an efficient and scalable solution for support in dealing with the ISO 26262 standard:

• Improved accessibility: Users receive quick and precise answers without having to search through long documents.
• Increased accuracy: By combining NLP technologies, relevant text passages are selected and context-sensitive answers are generated.
• Time savings: The automated provision of information reduces the time needed to search for specific content and makes it easier to apply the standard to operational processes.
• Broad applicability: The chatbot is easily accessible and understandable for users with different technical backgrounds.