网联协同驾驶避障场景识别及控制策略

Recognition and Control Strategy of Obstacle Avoidance Scenarios for Networked Cooperative Driving

随着无人驾驶技术的发展和无人物流在多行业的成功应用,产生了多种特定场景的无人物流技术方案。其中,最具代表性的场景有封闭厂区内的无人物流、机场无人物流、物流园区无人物流等。网联协同驾驶旨在结合5G通信技术低时延、大接入量的特点,将感知和运算部分从车端迁移到路侧和云端,并将无人驾驶道路感知部署在路侧,使用路侧提供的感知信息将无人驾驶控制器部署在云端,使用云端服务器进行路径规划和实时控制,大大降低车辆算力和感知的要求。使用网联协同驾驶的技术方案,需要根据道路信息定义避障场景,根据场景制定避障策略。笔者通过实验阐述了网联协同驾驶避障场景和避障策略,解决了网联协同驾驶车辆不搭载高精地图而导致的窄巷、弯道无法通行的问题。

With the development of unmanned technology and the successful application of unmanned logistics in many industries,a variety of unmanned logistics technology solutions for specific scenarios have emerged,among which the most representative scenarios are:unmanned logistics in closed plant area,airport unmanned logistics,express delivery unmanned logistics,etc.Networked cooperative driving aims to combine the characteristics of low delay and large access volume of 5g communication technology to migrate the sensing and computing part from the vehicle end to the roadside and cloud:deploy driverless road sensing on the roadside and use the sensing information provided by the roadside;Deploy the unmanned controller in the cloud and use the cloud server for path planning and real-time control.Greatly reduce the requirements of vehicle computing power and perception.Using the technical scheme of networked cooperative driving,need to define obstacle avoidance scenarios according to the road information and formulate obstacle avoidance strategies according to the scenarios.Through experiments,this paper expounds the commonly used obstacle avoidance scenes and obstacle avoidance strategies of networked cooperative driving,and solves the problem of impassability in narrow lanes and curves caused by networked cooperative driving vehicles not carrying high-precision maps.