NGSIM DATA PROCESSING

Robotics and Intelligent Vehicle Automation Lab (RIVAL)

• Built by Dong Chen, Zhaojian Li from Michigan State University
• Started on Oct.23, 2019, Lastly updated on Oct.25, 2019

Overview

This resposity aims at building car following dataset. We use a subset of NGSIM dataset, full dataset can be found here.

Motivation:

Next Generation Simulation (NGSIM) program collected vehicle trajectory data on southbound US 101 and Lankershim Boulevard in Los Angeles, CA, eastbound I-80 in Emeryville, CA and Peachtree Street in Atlanta, Georgia. We want to find the trajectories of vehicles and fit the data to our car-following model. This resposity aims at find sequences of vehicles and corresponding trajectory information where the vehicles inside the sequences follow each other without lane changing or extra vehicles enter the sequences during a time interval.

Part1. Lane Change Analysis

There are three subdataset. We use the dataset Data-0500-0515 of Next Generation Simulation (NGSIM) Interstate 80 Freeway Dataset. You can find the dataset analysis from the Attachments. The dataset shows like this.

Fig.1 Column 1-6 of Data-0500-0515

Fig.2 Column 7-12 of Data-0500-0515

Fig.3 Column 13-18 of Data-0500-0515

We can find the same vehicle (Vehicle_ID) occurs many time as the dataset shows the trajectory for each vehicle in frame (the frame rate is 15).

The code in data_process.m processes the trajectory information for each car. For each lane, there is a file lane_analysis_X.mat in the /lane_analysis/ folder.

Fig.4 Data in lane_analysis_2.mat

Columns are Vehicle_ID, Preceding, Forwarding, start_frame, end_frame, total_frames, start_y, end_y, lane_change_flag. That's the information related to each vehicle arranged in the order of entering time. Here we can see that the sequence of vehicles with Vehicle_ID 353--> 374--> 387 --> ... ---> 467 meets our requirement, as there is no lane changing event happening (lane_change_flag is 0). We can also use the trajectory plotting to prove that.

Part2. Trajectory Plotting

We can plot the trajectories of vehicles over time frames. The code in plot_trajectories.m does trajectory plotting.

Fig.5 trajectory_plotting

Part3. Data Saving

The code in save_data.m saves the vehicle and related information.

Cite

@misc{hajidavalloo2020mechanical,
      title={A Mechanical System Inspired Microscopic Traffic Model: Modeling, Analysis, and Validation}, 
      author={Mohammad R. Hajidavalloo and Zhaojian Li and Dong Chen and Ali Louati and Shuo Feng and Wubing B. Qin},
      year={2020},
      eprint={2012.02948},
      archivePrefix={arXiv},
      primaryClass={eess.SY}
}

Reference

[1]. Next Generation Simulation (NGSIM) Vehicle Trajectories and Supporting DataAutomobiles

[2]. NGISM-I-80-Trajectory-Animation