Improving the efficiency of software integration testing is a critical challenge in the automotive industry, particularly as Electronic Control Unit (ECU) architectures become increasingly complex. This paper addresses the automation of integration test script generation by leveraging Large Language Models (LLMs) with Retrieval Augmented Generation (RAG). Specifically, we target the phase in which test engineers translate natural language test case specifications into executable scripts for integration test environments containing hardware debug interfaces. To bridge the knowledge gap between LLMs and domain-specific test tool APIs, we construct a task-oriented vector store that incorporates both API manuals and supplemental, workflow-centric information. By combining these with prompts containing code prefixes, our method enables LLMs to generate robust and correct integration test scripts. We evaluated our approach on typical test scenarios reflecting industry practices for multi-core ECUs. While the test cases used were not directly taken from a specific development project, they closely mirror those routinely employed across numerous automotive ECU development initiatives. The proposed method successfully generated executable scripts for all cases and reduced total test execution man-hours by 43% compared to a realistic baseline of manual execution. These results demonstrate the practical benefit of context-enriched LLMs in accelerating specialized software engineering tasks within the automotive domain, and it also identifies remaining challenges in extending automation to broader aspects such as test coverage, maintainability, and seamless process integration.

The number of functions and the sizes of electronic controls and software systems in automobiles are increasing, as ADASAdvanced Driver Assistance Systems

Most automotive control systems have been developed by combining legacy Electronic Control Units (ECUs) with newly developed ECUs. We need to plan the product line to utilize as many legacy ECUs as possible to lower development costs of new vehicles. However, we must develop several new gateway ECUs per new vehicle because gateway ECUs depend on the placement of other ECUs and their combinations within the whole system. The reusable gateway ECUs are needed to lower development cost. However, there is a problem in that the worst case execution time (WCET) of relay processing may be impaired in typical methods that are synchronized with the communication cycle of time-triggered networks. Therefore, we propose a Dynamic Activation Timing Configuration (DATC) for product line development, which can dynamically calculate the activation timing of relay processing from the information on message routing, slot assignment, and time-triggered network. The gateway ECUs with DATC can be dynamically adjusted to new vehicles and improve the WCET of relay processing.

This paper describes a novel approach to detect variability in a software product line from its change history such that software changes are converted to vectors and a factor analysis is applied. To show the applicability of our approach, we conducted experimental applications using a software repository of automotive engine control software. As a result of the experiments, variability is detected from the change history of products.