Workflow
交通视觉语言模型
icon
Search documents
高德发布鹰眼守护预警系统,实测重大异常事件预警准确率达90%
Feng Huang Wang· 2025-09-28 11:16
Core Viewpoint - The "Eagle Eye Guardian" warning system, developed by the China Academy of Safety Science and Technology in collaboration with Amap, aims to enhance road safety by providing proactive alerts to drivers about potential dangers beyond their immediate visibility, especially during the busy travel period of the Mid-Autumn Festival and National Day holiday [1][2]. Group 1: System Overview - The "Eagle Eye Guardian" system operates through three main stages: perception, judgment, and warning. It detects sudden deceleration or braking of vehicles ahead and uploads this information anonymously to a cloud server for real-time analysis [2]. - The system utilizes a traffic visual language model called TrafficVLM, which combines traffic twin restoration and visual language modeling to provide real-time optimal decision-making for users [2]. Group 2: Testing and Performance - In August, the technology underwent real vehicle testing at the Shandong Intelligent Connected Highway Testing Base, demonstrating a 90% accuracy rate in warning drivers of significant abnormal events, such as multiple vehicles suddenly braking at speeds over 80 km/h, with alerts received within seconds and covering nearly 1 kilometer [2]. Group 3: Coverage and Scenarios - Amap has integrated the "Eagle Eye Guardian" warning system into 17 common travel scenarios, including major abnormal events, vehicles approaching around curves, intersections without traffic lights, sudden braking ahead, slow-moving vehicles, fast-approaching vehicles from behind, and vehicles entering or exiting junctions [3].