Recognising wild animals on roads: multisensor systems for accident avoidance

Abstract. Wildlife-related traffic accidents represent a persistent hazard on rural roads in Germany and beyond. Current electronic wildlife warning systems typically monitor only very short distances and therefore cannot provide large-area coverage. This paper presents a novel multisensor approach that integrates radar and infrared (IR) technology into existing roadside delineators. Due to regulatory requirements, delineators are placed at intervals of 50 m on German country roads. Integrating sensors into these delineators thus provides a uniform infrastructure that can be utilised. The radial extension of the sensor range allows a monitoring zone to be formed along the road. We evaluate thermal infrared arrays and high-resolution 60 GHz radar sensors for range, resolution and robustness under varying environmental conditions. Field measurements in wildlife parks demonstrate that the system can reliably detect deer at distances of up to 30 m and evaluate their moving speed as well. Challenges such as ambient temperature effects, optical dispersion in IR detection and resolution limits are discussed. The results highlight the potential of multisensor systems to reduce wildlife accidents and improve road safety.