Modeling Wildlife Accident Risk with Gaussian Mixture Models

Traffic accidents involving wildlife pose a widespread problem globally, harming both humans and nature. These incidents often result in heavy vehicle damage, leading to expensive repairs and insurance claims. To mitigate these accidents, efforts are underway to understand wildlife populations near high-risk roads better and implement preventive measures such as visual or audible wildlife warning devices. To prevent wildlife accidents, high-risk areas must be identified first. In this work, we propose a model that predicts dangerous areas based on animal sightings and apply it to two road segments in Austria.