正文 · AI 翻译
IT Home reported on November 8 that on November 5, 2026, Sony (China) Co., Ltd. (“Sony”) made the global debut of its “vision-language model-based driver takeover hazard warning” mobility technology at the 9th China International Import Expo.
According to the introduction, when road conditions approach or exceed the system’s operating limits, clear takeover prompts help drivers understand the situation and prepare to respond. The technology uses a vision-language model (VLM) to analyze images of the road ahead of the vehicle and determine whether conditions are approaching or exceeding the intelligent driving system’s operational design domain (ODD), providing drivers with tiered prompts.
IT Home noted that the takeover prompts have three levels: when conditions approach the boundary, they remind the driver to prepare to take over; when conditions exceed the operating range but are not yet urgent, they guide the driver to take over proactively; and in high-risk situations such as sudden obstacles or perception failures, they prompt an emergency takeover.
During development, the team used a digital twin simulation environment to generate data for conditions including daytime, nighttime, heavy rain and high-speed driving, for model training and validation. The technology has now undergone real-road testing with Tsinghua University’s School of Vehicle and Mobility in Xiangcheng District, Suzhou, and can support automakers and intelligent driving solution providers in deciding when drivers should take over and designing human-machine interaction.