Traffic forecasting with Virtual Induction Loops - SUMO simulation dataset

This repository is associated with my doctoral dissertation titled, "Smartphone based applications for Road Traffic Telematics". In particular this repository serves as the basis of Chapter 7 titled, "Traffic forecasting with Virtual Induction Loops (VIL)". The basic idea is to validate a traffic forecasting system which uses machine learning techniques on the simulation of real traffic flows on a real intersection in the City of Turin. This dataset contains simulation output from SUMO software for 56 real days between the months of October-2017 to April-2018. Details about these days are available in my thesis. For each day, 3 output files are available. Here is the description and naming convention: M1_100seed_100pr_dump.csv (This is the data dump file from SUMO. It contains flows of every single vehicle that was simulated. Naming convention is day_seed_vilPenetrationRate_dump.csv) M1_100seed_ilNorth_100pr.xml (This is the output from a simulated induction loop for Northbound traffic. Naming convention is day_seed_ilNorth_vilPenetrationRate.xml) M1_100seed_ilSouth_100pr.xml (This is the output from a simulated induction loop for Southbound traffic. Naming convention is day_seed_ilSouth_vilPenetrationRate.xml) For further details, please refer to my thesis.本仓库关联于笔者题为《智能手机在道路交通远程信息服务中的应用》的博士学位论文。具体而言，本仓库为第7章《基于虚拟感应线圈（Virtual Induction Loops, VIL）的交通预测》的核心支撑材料。本研究的核心目标为，针对意大利都灵市某真实交叉口的真实交通流仿真场景，验证一套采用机器学习技术的交通预测系统。本数据集包含SUMO软件于2017年10月至2018年4月间的56个真实日期的仿真输出结果，各仿真日期的详细信息可参阅笔者的博士论文。每日对应3个输出文件，其说明与命名规范如下：1. M1_100seed_100pr_dump.csv：该文件为SUMO导出的数据转储文件，包含本次仿真中所有模拟车辆的行驶流量数据，命名规则为日期标识_随机种子_VIL渗透率_dump.csv（原文格式为day_seed_vilPenetrationRate_dump.csv）；2. M1_100seed_ilNorth_100pr.xml：该文件为北向交通的仿真感应线圈输出结果，命名规则为日期标识_随机种子_ilNorth_VIL渗透率.xml（原文格式为day_seed_ilNorth_vilPenetrationRate.xml）；3. M1_100seed_ilSouth_100pr.xml：该文件为南向交通的仿真感应线圈输出结果，命名规则为日期标识_随机种子_ilSouth_VIL渗透率.xml（原文格式为day_seed_ilSouth_vilPenetrationRate.xml）。如需获取更多细节，请参阅笔者的博士学位论文。