信用分配下智能网联车轨迹数据价值评价方法  

Trajectory Data Value Evaluation Algorithm Based on Credit Assignment

作　　者：陈越 尹嘉诚[1,2] 曹鹏 CHEN Yue;YIN Jia-cheng;CAO Peng(School of Transportation and Logistics,Southwest Jiaotong University,Chengdu Sichuan 611756,China;Yibin Research Institute,Southwest Jiaotong University,Yibin Sichuan 644000,China)

机构地区：[1]西南交通大学交通运输与物流学院,四川成都611756 [2]西南交通大学宜宾研究院,四川宜宾644000

出　　处：《计算机仿真》2024年第12期183-188,共6页Computer Simulation

基　　金：四川省自然科学基金(2022NSFSC0476);国家自然科学基金(61903313)。

摘　　要：智能网联车能够采集高精度轨迹数据,汇集多辆智能网联车轨迹数据,能够提高交通系统的表征效果。当下多以采集数据量多少作为评价标准,此方法受到时空流量分布不均、采集数据量边际效用递减等影响而不准确。为解决上述问题,提出了一种基于信用分配的评价方法。将轨迹数据根据划分时间段轨迹重构,计算当前时段价值;基于MLP预测最终时段价值,与前次预测价值比较,得出历次预测结果;依据采集轨迹数据价值权重,结合预测变化趋势,得到当前时段每辆智能网联车轨迹数据价值;根据不同时段、价值权重,求出全时域智能网联车轨迹数据价值。对比试验与敏感性分析结果表明,本方法的皮尔逊相关系数显著高于传统方法,说明本方法能更好地评价智能网联车轨迹数据价值。Intelligent networked vehicles can collect high-precision trajectory data,and the collection of multiple intelligent networked vehicle trajectory data can improve the characterization effect of the traffic system.Most of the current evaluation criteria are based on the amount of collected data,and this method is affected by the uneven distribution of spatial and temporal traffic and the diminishing marginal utility of the collected data volume,which is inaccurate.For this reason,this paper proposes an evaluation method based on credit distribution.The trajectory data are reconstructed according to the divided time period trajectory,and the current time period value is calculated;the final time period value is predicted based on MLP,the previous predicted value is compared with the previous predicted value,and the historical prediction result is obtained.Based on the value weight of the collected trajectory data,combined with the predicted change trend,the value of each intelligent networked vehicle trajectory data in the current time period is obtained.According to different time periods and value weights,the value of the full-time domain intelligent networked vehicle trajectory data is derived.The results of the comparison test and sensitivity analysis show that the Pearson correlation coefficient of this method is significantly higher than that of the traditional method,which indicates that this method can better evaluate the value of intelligent networked vehicle trajectory data.

关 键 词：智能交通 交通数据评价 信用分配 智能网联车 轨迹重构 神经网络 

分 类 号：U268.6[机械工程—车辆工程]