Design and Development of a Data-Driven Animal Collision Avoidance System

2025-01-5039

06/10/2025

Features
Event
Automotive Technical Papers
Authors Abstract
Content
A collision avoidance system is an advanced driver assistance system (ADAS) designed to prevent vehicle collisions. The existing system is designed primarily to detect and prevent collisions with humans, vehicles, or other objects on the road, but not specifically designed to address collisions involving animals. When it comes to animals, the collision avoidance rate is low, which is due to low animal detection accuracy, high false positives/negatives and unsuitable methods of mitigating collisions, hence, an enhancement of the system is required. Humans are warned off the road during the approach of a vehicle through the horn sound. In ADAS systems, the vehicle stops as the human behavior is quite predictable on sight or approach of a speeding vehicle, but this is ineffective on animals. Not all animals can hear the honk due to their different hearing ranges, and the unpredictability of their behaviors makes it impossible to predict their next move to avoid an accident; this leads ultimately to a collision even when equipped with the most advanced collision avoidance systems.
There needs to be a more preventive approach to keep the animals that are likely to collide, off the road. It can be done with enhancement of the collision avoidance system. The animal collision avoidance system (ACAS) uses ADAS sensor suite and GPS inputs to detect and predict the behavior of an approaching animal. Further, the ACAS calculates the likelihood of collision and reacts accordingly to warn off the animal either through a threat warning in its hearing range or by adapting the vehicle to not collide with the animal.
Meta TagsDetails
DOI
https://doi.org/10.4271/2025-01-5039
Pages
15
Citation
Arckja, W., "Design and Development of a Data-Driven Animal Collision Avoidance System," SAE Technical Paper 2025-01-5039, 2025, https://doi.org/10.4271/2025-01-5039.
Additional Details
Publisher
Published
Jun 10
Product Code
2025-01-5039
Content Type
Technical Paper
Language
English