Revolutionizing Requirements Engineering in Automotive E/E System Development with Generative AI

August 28, 2024
Generative AIRequirements Engineering

Generative AI has the potential to transform businesses by automating several repetitive tasks and significantly enhancing efficiency and productivity. According to a recent McKinsey Report on Gen AI uses in Automotive R&D, automotive OEMs & suppliers have already achieved productivity gains ranging from 20% to 70% by introducing Generative AI based copilots and tools across various activities such as requirements engineering, software testing, validation and homologation.

In the complex world of automotive Software Defined Vehicle (SDV) and Electric & Electronic (E/E) system development, Requirements Engineering (RE) is an important process. It ensures that all system specifications meet the desired functionality, safety, security and regulatory standards. However, a lot of manual tasks in RE, ensuring stringent ASPICE compliance, the complexity and volume of requirements can pose significant challenges.

AI in Requirements Engineering

Generative AI offers a transformative solution to enhance the efficiency and accuracy of RE processes to ensure compliance with ASPICE. With advancements in AI, especially Natural Language Processing (NLP), AI algorithms are now capable of optimizing many time consuming tasks in RE. A recent research on applying AI methods in RE has presented a set of standardized process steps which could be used to systematically identify AI application opportunities in RE. Let’s explore the typical activities and problems encountered in the RE process and how generative AI can address them.

Requirements Elicitation

The first step in RE is requirements elicitation and documentation. This involves gathering information from stakeholders to define what the system should do. This stage is met with challenges such as ambiguity, where stakeholders may provide unclear or incomplete information, and miscommunication, which can lead to conflicting requirements from different stakeholders. Additionally, handling a large volume of detailed information can be overwhelming.

Generative AI can significantly enhance the elicitation and documentation process by employing Natural Language Processing (NLP) to interpret and translate stakeholder input into formal requirements, reducing ambiguity and miscommunication. AI-driven chatbots can conduct initial stakeholder interviews, ensuring comprehensive and consistent information gathering, while AI tools can summarize lengthy stakeholder discussions into concise requirement statements, making documentation more manageable.

Requirements Analysis

Next is the requirements analysis phase, which involves evaluating requirements to ensure they are complete, feasible, and aligned with business goals. Challenges in this stage include ensuring all requirements are consistent with each other, assessing whether the requirements can be realistically achieved within constraints, classifying & allocating the requirements to different teams and determining the importance of each requirement, which can be subjective and contentious.

Generative AI can help by automatically detecting and highlighting inconsistencies among requirements, ensuring coherence. Using historical data and predictive analytics, AI can evaluate the feasibility of requirements against technical and resource constraints. Furthermore, AI can help to classify & allocate requirements based on set parameters and prioritize requirements based on criteria such as stakeholder impact, cost, and risk, making the process more objective and data-driven.

Requirements Specification

In the requirements specification phase, the challenge is to document the requirements in a detailed and structured format. This involves refining and writing requirements at different abstraction levels in a clear, unambiguous, and testable manner, ensuring traceability by linking every requirement to its origin and rationale, and maintaining the documentation up-to-date as requirements evolve.

Generative AI can ensure that requirements are written in clear, consistent, and unambiguous language as per predefined guidelines (e.g. INCOSE, EARS, etc.), reducing misinterpretation. It can also automatically link requirements to their sources, related documentation, and rationales, enhancing traceability. AI-driven tools can maintain and update requirement documents for each new iteration, ensuring they reflect the latest changes and developments.

Requirements Review

Finally, the requirements review phase involves validating requirements through stakeholder review and feedback. Ensuring that all requirements have been captured and reviewed, making sure all stakeholders correctly interpret the requirements, and achieving agreement among stakeholders can be challenging.

Generative AI can streamline the review process by pre-checking requirements for completeness and compliance with predefined standards before stakeholder review. AI can analyze feedback and identify common themes or concerns, facilitating more focused and productive review sessions. Additionally, AI can provide data-driven insights and recommendations to help stakeholders reach agreement on contentious requirements.

Benefits of Generative AI in RE

Generative AI represents a powerful tool in the arsenal of automotive E/E system development, offering significant enhancements across all stages of the requirements engineering process. By addressing the challenges of elicitation, analysis, specification, and review, AI enables teams to produce higher quality requirements more efficiently and effectively and ensures ASPICE compliance with end-to-end traceability. As the automotive industry continues to evolve, embracing AI-driven tools will be crucial in meeting the ever-increasing demands for innovation, safety, and reliability.

Leverage Generative AI in RE with Teratics

At Teratics, we offer Generative AI based Assistants tailor-made for automotive E/E system development that includes requirements engineering processes to reduce cost, increase productivity & quality and ensure compliance with industry standards.