Capturing vehicular space headway using low-cost LIDAR and processing through ARIMA prediction modeling

TL;DR

This paper presents a low-cost, integrated system using compact sensors and ARIMA modeling to accurately capture and filter vehicular space headway data for traffic analysis.

Contribution

The paper introduces a novel low-cost sensor system combined with ARIMA-based filtering for precise vehicular headway data collection.

Findings

Successfully captures spatial headway and speed data at low cost

Effective filtering reduces data outliers significantly

Enables detailed car-following and speed-density analysis

Abstract

The paper proposes a low-cost system to capture spatial vehicle headway data and process the raw data by filtering outliers using a novel filtering process. Multiple sensors and modules are integrated to form the system. The sensors used are compact, lightweight, low-cost and have low power consumption. A single beam 1-Dimensional Light Detection and Ranging (LIDAR) was used for capturing the space headway data, a Global Positioning System (GPS) to map each data point with a timestamp and position and also a camera to capture video data with an overlay of date, time, distance and speed in real-time. The filtering technique utilizes the Autoregressive Integrated Moving Average (ARIMA) prediction modeling and mean-filtering. The data captured is stored in a Raspberry Pi module. The data is later processed by using the filtering technique to obtain the least outliers. The overall system has enabled to capture spatial headway data and speed of the vehicle at a very low cost and the data obtained can be used for car-following model analysis and speed-density analysis.