VEDAS Project

VEhicle DAta Stream Mining (VEDAS) Project: An Experimental System for Mobile and Distributed Data Stream Mining


Figure 1. VEDAS application architecture.

The Vehicle Data Stream Mining System (VEDAS) is a mobile and distributed data stream mining/monitoring application that taps into the continuous stream of data generated by most modern vehicles. It allows continuous on-board monitoring of the data streams generated by a moving vehicles, identifying the emerging patterns, and reporting them to a remote control center over low-bandwidth wireless network connection. The system offers many exciting possibilities such as real-time on-board health monitoring, drunk-driving detection, driver characterizations, and security related applications for commercial fleet management. The system is comprised of three main components:

1. Hardware interface for the on-board OBD-II data bus that couples with our software:


Figure 2. VEDAS Hardware interface.

This module was developed in-house for interfacing with the on-board OBD-II data bus. It also contains a GPS module.

2. On-board data stream management and mining module that runs from a PDA:

The VEDAS system offers a communication system and run time environment for performing on-board data analysis and management. The on-board PDA-based module monitors, manages the data stream and triggers action when unusual activities are observed. The on-board module connects to the desk-top-based remote control station through low-bandwidth wireless network.The system allows the fleet managers to monitor and model vehicle behavior remotely without necessarily downloading all the data to the central moitoring station over the expensive wireless connection.

3. Remote desktop-based control station for fleet managers:

Figure 3. Graphical user interface of the VEDAS Control Station

The VEDAS control station supports the following main operations:

1) Interacting with the on-board module for remote management, monitoring, and mining of vehicle data streams.

2) Interactive statistical data analysis

3) Interactive online PCA-based vehicle health regime monitoring.

4) Driving characterization.

A commercial version of this system is now developed by AGNIK, LLC.