Intelligent test management is essential for increasing efficiency, transparency, and regulatory compliance in modern battery testing environments.
In this thesis, a conversational AI agent will be developed as a front end to an prototype test management and scheduling framework. The agent will support engineers in defining, optimizing, and evaluating test campaigns by combining static information (e.g. regulations, testing standards, use-case descriptions) with dynamic information (e.g. bench availability, model availability, current test load).
-
Conducting a literature and state-of-the-art review on AI-based assistants, test scheduling, and regulatory frameworks with or without focus for battery testing
-
Analysing relevant regulatory documents, testing standards, and project-specific user requirements, and translating them into machine-readable rules and constraints for test planning
-
Designing the system architecture of an AI chat-bot that acts as a natural-language front end for a test scheduler
-
Implementing the AI agent with preference of using large language models (LLMs) including interfaces to data sources such as test bench calendars, model repositories, and requirement databases
-
Developing algorithms or workflows for test recommendation and automatic scheduling under static and dynamic boundary conditions (standards, priorities, availability, cost or time constraints)
-
Comparing the AI-based results with a baseline conventional optimisation-based scheduler and defining quantitative evaluation criteria (e.g. utilisation, waiting time, deadline adherence, compliance with constraints)
-
Running a case study on battery testing for technical feasibility assessment for the desired purposes
-
Documenting the results in a scientific manner and preparing the Master thesis and an internal presentation
