Using Technology to Reduce Wildlife-Vehicle Collisions Symposium

The team working on iMOVE’s Large Animal Activated Roadside Monitoring and Alert System is presenting project outcomes next week at the Using Technology to Reduce Wildlife-Vehicle Collisions Symposium.

Participants on that iMOVe project are Department of Transport and Main Roads (Queensland), University of Sydney/ Australian Centre for Field Robotics, and Queensland University of Technology/ CARRS-Q. Sponsoring the event is another iMOVE participant, Transport for NSW, along with the Australasian Network for Ecology and Transportation (ANET), and the Environment Institute of Australia and New Zealand (EIANZ).

Topics being presented at the event include:

• Animal detection and identification systems
•  Animal deterrent systems (e.g. virtual fences)
•  In-car and roadside warning systems for motorists
• Vehicle automation

Paper and abstract

The paper being presented at the symposium is titled Enhancing Road Safety and Wildlife Conservation in Regional Australia: The Development of a Novel Large Animal Activated Roadside Monitoring and Alert System, and has been co-authored by Kunming Li, Mao Shan, Sebastien Glaser, Ioni Lewis, Stewart Worrall, Mohammed Elhenawy, Sebastien Demmel, and Xiaomeng Li.

Wildlife-vehicle collisions (WVCs) present a significant threat to wildlife conservation and human safety, particularly in wildlife-rich areas like regional Australia. Conventional mitigation measures, such as wildlife crossings and fencing, are often limited by geographical and financial constraints. Innovative machine-learning-based approaches have emerged as promising solutions for animal detection; however, these approaches face challenges in detecting species where there is insufficient existing data for model training.

This collaborative study, conducted by the University of Sydney, Queensland University of Technology, and Queensland Department of Transport and Main Roads, develops a novel large animal activated roadside monitoring and alert (LAARMA) system. The system monitors roadside animals via cost-effective perception sensor suite and alerts motorists to the safety hazard in real time via purpose-designed messaging displayed on roadside variable message signs (VMSs).

An innovative feature of the animal detection is its self-supervised learning pipeline, which enables the system to automatically label real-world animal data collected during field operations, continually improving its accuracy and reliability without extensive human supervision. Prior to their use in the on-road trial of the technology, a series of messaging concepts were developed, concept-tested via qualitative focus groups and evaluated via a large-scale online survey of drivers’ responses to messaging. Such aspects were underpinned by robust conceptual and methodological approaches to message design and evaluation; namely, the Step approach to Message Design and Testing (SatMDT).

Two selected messages were evaluated on 50 participants in a 40-minute drive in a driving simulator study to evaluate the fine response to the developed messaging strategy. The overall system’s performance and road safety outcomes are evaluated using field data from the real-world system deployment at a site in Far North Queensland.

Register for the event

The Using Technology to Reduce Wildlife-Vehicle Collisions Symposium will take place on 21 May 2024, at the University of Technology Aerial Function Centre, in Sydney.

Registration is required for both in-person and online participation. Rates for on-line participation are the same as in-person due to increased audio-visual associated with running an interactive online and in-person event.

The link to register, and for additional information, is at:Symposium | Using technology to reduce wildlife-vehicle collisions