摘要
自动驾驶汽车技术近年来得到了广泛研究与快速发展,但在复杂的交通场景下,自动驾驶汽车面对突然出现的行驶车辆并不能及时避让。针对此问题,基于区块链技术提出智能网联车队的协同轨迹预测系统,智能网联车队中的各个节点与路边基础设施通过长短时记忆网络(LSTM)模型对周边车辆的运动轨迹进行预判,并将得到的结果进行分享,利用区块链技术,智能网联车队与路边基础设施可以对其接收到的信息进行评分,并将汇总后的评分以区块的形式加入存储信誉评分的区块链中。通过该评分,智能网联车队中的车辆可以根据车队中其他节点的信誉值来判断其是否可信,低信誉值节点传来的信息将不予理睬,从而实现了协同驾驶。实验分析表明,所提LSTM模型能够较为准确地预测周边车辆5 s内的行驶轨迹,而所提的系统在提升智能网联车队的行驶安全上起到了明显的效果。
Autonomous vehicles have been extensively researched and developed rapidly in recent years.However,these vehicles may be not able to avoid collision with vehicles that appear suddenly.To solve this problem,a cooperative trajectory prediction system for intelligent and automated connected convoy based on blockchain was proposed.In the proposed system,each node in the intelligent and automated connected convoy and roadside infrastructure predict the trajectory of vehicles around them with the LSTM model,and they share the information they get.Using blockchain,the intelligent and automated connected convoy and roadside infrastructure rate the received messages.The summarized rating will be stored as a block in the blockchain that stores the trust values.Vehicles in the intelligent and automated connected convoy are able to judge whether the other nodes in the Conroy are credible according to their trust value,and messages from the low trust value sender will be ignored,thus achieving cooperative driving.Experiment shows that the proposed system plays a significant role in improving driving safety of the intelligent connected platoon.
作者
王仁生
吴天昊
张琳
朱孔林
WANG Rensheng;WU Tianhao;ZHANG Lin;ZHU Konglin(School of Information and Communication Engineering,Beijing University of Posts and Telecommunication,Beijing 100876,China)
出处
《网络与信息安全学报》
2020年第3期78-87,共10页
Chinese Journal of Network and Information Security
基金
国家重点研发计划基金(2016YFB0100902)。
关键词
智能网联车队
轨迹预测
长短时记忆网络
信任管理机制
区块链
intelligent and automated connected convoy
trajectory prediction
LSTM
trust management
blockchain
作者简介
通信作者:王仁生(1994−),男,山东烟台人,北京邮电大学硕士生,主要研究方向为智能网联汽车。wangrensheng@bupt.edu.cn;吴天昊(1992−),男,山东济南人,北京邮电大学博士生,主要研究方向为C-V2X通信技术及应用;张琳(1974−),男,上海人,博士,北京邮电大学教授、博士生导师,主要研究方向为超宽带生物雷达成像及生命信号检测、网络信息处理、车联网;朱孔林(1985−),男,山东临沂人,博士,北京邮电大学副教授,主要研究方向为智能网联汽车。
