车路协同路侧感知融合方法的研究

随着智能网联汽车技术和产业的不断发展,智能网联汽车逐渐成为人们交通出行的选择之一。但受智能网联汽车自身环境感知系统对特定道路交通场景信息处理的局限,无法实现在所有行驶工况下安全高效的运行,其需车路协同路侧感知技术的辅助方能更安全高效的运行。海量的车路协同感知数据是城市道路和高速公路车路协同、运行分析和科学管理的宝藏,理解和分析这些数据是车路协同路侧感知融合的关键。面对车路协同路侧多传感器的不同数据,如何高效准确地挖掘和提取雷达、视频在不同时间、不同空间维度的数据,实现对重点交通场景(如视野盲区、急转弯道、隧道、桥梁)和交通事件、环境、设施安全等的雷达、视频数据进行快速融合检测、识别与检索,通过蜂窝车联网C-V2X网络在一定时延范围内有效地将路侧感知融合结果数据发送给智能网联汽车,确保其安全高效的行驶,是面向智能网联汽车辅助驾驶的车路协同路侧感知融合的关键问题。基于智能网联汽车其自身环境感知能力,对道路智能基础设施感知网络中的多传感器融合方法进行研究分析,提出了基于误差方差的多传感器融合算法,与非智能道路相比,其效率更高,更加智能化,可有效解决道路交通运行环境中存在的常见问题,为人们提供更加安全、高效、优质的交通出行服务。

Vehicle Road Cooperative Roadside Perception Fusion Method

With the continuous development of intelligent connected vehicle technology and industry,intelligent connected vehicles have gradually become one of the choices of transportation.However,due to the limitations of the intelligent connected vehicle’s own environmental perception system on the information processing of specific road traffic scenes,it can not achieve safe and efficient operation under all driving conditions.It needs the assistance of vehicle road cooperative roadside perception technology to operate more safely and efficiently.Massive vehicle road collaborative perception data is a treasure of vehicle road collaboration,operation analysis and scientific management of urban roads and expressways.Understanding and analyzing these data is the key to vehicle road collaborative roadside perception fusion.Facing the different data of vehicle road cooperation roadside multi-sensor,how to efficiently and accurately mine and extract the data of radar and video in different time and different spatial dimensions so as to realize the rapid fusion detection,identification and retrieval of radar and video data of key traffic scenes(such as blind area of field of vision,sharp turns,tunnels and bridges)and traffic events,environment and facility safety,and effectively send the roadside perception fusion result data to the intelligent connected vehicle within a certain time delay through the cellular car network C-V2 X network,to ensure its safe and efficient driving,is the key problem of vehicle road cooperative roadside perception fusion for intelligent connected vehicle assisted driving.Based on its own environmental perception ability,the multi-sensor fusion method in the road intelligent infrastructure perception network is studied and analyzed,and a multi-sensor fusion algorithm based on error variance is proposed.Compared with non intelligent roads,it is more efficient and intelligent,and can effectively solve the common problems existing in the road traffic operation environment and provide people with more safe,efficient high quality transportation service.