Non-invasive vehicle classification solution using tracking radar

This PhD project aims to research and develop new algorithms that can process the raw data generated by the current RT4 radar and assign an accurate Austroads classification to all detected vehicles, in addition to the speed measurement and length returned by the current RT4 algorithms.

Participants

• Sensys Gatso Group
• University of Melbourne

Project background

One of the local developments carried out by Sensys Gatso Australia (SGA) has included the GT20-S Automatic Number Plate Recognition (ANPR) system. This has primarily been employed for heavy vehicle enforcement projects, where the industry has expressed a desire to find an alternative non-invasive vehicle classification system

in accordance with the Austroads standard. The Austroads classification system is based around the number of axles, their groupings and the spacing between these groups. There is a total of 12 classes in the scheme, starting from a motorbike up to a road train.

SGA has developed a mobile speed camera system based around the RT4 tracking radar from SGG. As a result of the extensive testing and product refinement that has taken place over the past 2 years, it has become apparent that the RT4 algorithms could be modified to enable Austroads classification.

This project therefore aims to research and develop new algorithms that can process the raw data generated by the current RT4 radar and assign an accurate Austroads classification to all detected vehicles, in addition to the speed measurement and length returned by the current RT4 algorithms.

Additional benefits of the algorithms may include, but not be limited to overloaded vehicle detection (mass distribution) and vehicle height and width measurement.

Whilst this project proposes to develop algorithms suitable for running on existing tracking radar technology, the modular nature of the algorithms developed will enable them to be applied on multiple (alternate) radar hardware configurations.

It is intended to employ signal processing methods, including machine learning and/or deep learning tools, in order to successfully complete this project.

The output of this project can benefit the community by being utilised to provide increased safety to road users. This can be achieved where the technology is used to monitor the time vehicles are on the road (driver fatigue), whether vehicles are overloaded and pose a danger to other road users and if oversized vehicles are travelling on roads not suited to accommodate their specific characteristics.

Indeed, the statistical data gathered would enhance the road amenities project plan to reduce traffic congestion due to heavy vehicle and traffic intermixing.

Project objectives

The objectives of this project are:

1. Realisation of an algorithmic solution using the current radar technology allowing to provide a non-invasive axle-based vehicle classification which covers the full Austroads scheme.
2. Make this solution modular so it can be implemented on different hardware platforms, including the current and future radar technologies in development by SGG and SGA.

As agreed by all parties, if the project progress enables extra time to include an additional feature implementation of the project could be extended to:

3. Be applied overseas by modifying it so that it also satisfies the US FHWA Scheme F classification.
4. Investigate the implementation of the following additional features:

• Overloaded vehicle detection
• Vehicle height and width measurement

About Sensys Gatso Group

Sensys Gatso Group (SGG) is a traffic enforcement system manufacturer based in Sweden and the Netherlands. To date, SGG’s primary focus has been developing products for the enforcement market, rather than for traffic data collection purposes.

The Australian branch of the company takes off-the-shelf products from SGG and other suppliers and customises them to meet local requirements. Some of these local developments have included back office interfaces, incident and log file formatting, operator interfaces, and integrating secondary speed enforcement.

Update: April 2022, Profile of researcher, Victor Deville

We’ve interviewed Victor about his research, and about his PhD journey. Read his profile at Victor Deville – iMOVE PhD student.

Please note …

Ongoing, this page will be a living record of this project. As it continues, matures, hits milestones, etc., we’ll add information, links, images, interviews and more. Watch this space!