A Review on Interaction Design for In-Vehicle Warning Systems: Multimodal Fusion, User Adaptation, and Ethical-Legal Frameworks

As intelligent driving enters a transitional phase, in-vehicle warning systems are central to human-machine interaction, addressing challenges in functional upgrades and complex scenario adaptation. This review synthesizes interdisciplinary progress, focusing on three themes: (1) multimodal sensory fusion and neural computing mechanisms, including physiological-cognitive foundations, brain-computer interfaces (BCI), and generative AI applications; (2) user-centered behavior modeling and interaction adaptation, covering driving characteristics, strategies for diverse groups, and cultural differences; (3) ethical-legal frameworks for algorithm fairness, shared responsibility, bias correction, privacy protection, and region-specific regulations. Findings highlight research gaps, such as multimodal coordination mechanisms and cross-cultural empirical evidence. Key challenges include limited technical robustness in dynamic environments, unstructured user adaptation methods, and delays in ethical-legal integration. Future research should develop a unified “physiological-cognitive-technical-ethical” framework to support conceptual and practical advancements in autonomous driving implementation.