01659757

Component

Global Position System (GPS)-enabled smartphones provide new opportunities for location-based services and traffic estimation. Using the inbuilt GPS sensor with a smartphone for navigational purposes demands a considerable amount of battery usage and wireless communication with the remote server of the service providers. This creates an issue when the smartphone is running low on battery charge, or when there is no stable data connectivity. At the same time, continuous monitoring of end users via such sensors also imposes privacy/security concerns. In this paper, the authors present a self-adaptive sampling (SAS) method for GPS based data collection in real time. It uses the estimated traffic flow state from the individual trajectory to adjust the sampling rate of GPS data. The methodology uses a Hidden Markov Model (HMM) classifier to classify individual data points in to one of four basic traffic flow states (free flow, stopped, acceleration, and deceleration). The identification of traffic state of flow is used in real time for varying the sampling rate of the GPS data points. The authors found that the total amount of data could be reduced 65-85%, while keeping the most critical data points. The accuracy of the traffic state estimation model was around 75%, while the reduced data showed promising result in traffic modelling application and privacy protection as well.

This paper was sponsored by TRB committee ABJ70 Standing Committee on Artificial Intelligence and Advanced Computing Applications.

18-06345

English

6p

2018

Transportation Research Board 97th Annual Meeting

Digital/other

Figures; References; Tables

Data collection; Global Positioning System; Markov processes; Privacy; Sampling; Smartphones; Traffic flow

Data and Information Technology; Highways

Transportation Research Board Annual Meeting 2018 Paper #18-06345

TRIS, TRB, ATRI

Jan 8 2018 11:38AM