Modern automotive infotainment systems and digital instrument clusters require rigorous testing to ensure reliability and compliance with industry standards. Traditional testing methods are labor-intensive and lack scalability. This research proposes a phased, vision-based approach using image processing, machine learning (ML), and computer vision (CV) to enhance testing automation and accuracy.
Phase 1 employs open-loop testing using an NVIDIA Jetson Nano with camera modules to capture visual data from infotainment and instrument cluster displays. This phase focuses on the static analysis of visual elements, screen transitions, animations, and error messages. Feature extraction and pattern recognition techniques identify interface behaviors and detect anomalies without a feedback loop. This research will focus in depth on Phase 1, providing a detailed evaluation of its implementation and effectiveness. Phase 2 extends the testing environment with a hardware-in-the-loop (HIL) system integrated with the Jetson Nano setup for closed-loop testing. The HIL system simulates dynamic user interactions, while CV methods assess user inputs and system responsiveness. This paper will propose the architecture for Phase 2, setting the foundation for its future development as a part of ongoing research.
By integrating image processing, ML, and CV techniques, this framework addresses challenges in high cognitive load management, multimodal interaction validation, and system integration. The proposed approach reduces manual testing efforts, improves repeatability, and enhances accuracy. Ultimately, it contributes to refining next-Generation infotainment and digital cluster designs, ensuring improved safety, usability, and alignment with advanced driver assistance systems (ADAS).