Deployed system2021 to 2025ICOET 2023
RADIS Wildlife Detection System
Edge computer vision and real-time driver warnings
RADIS combines on-device wildlife detection, trajectory estimation, and RF-linked warning signs in a remotely powered roadside system. The system was field-deployed near Reno and presented at ICOET.
- Edge AI
- Computer vision
- Embedded systems
Outputs
Deployment
Roadside field deployment
Jetson-based detection, renewable power, RF communications, and driver-warning hardware operated as one field system.
Presentation
RADIS: Rapid Animal Detection and Identification System for Driver Warning
View presentationPress
Inside Edition
Video coverage of the roadside wildlife-warning system.
Watch coveragePress
GeekWire and KIRO 7
Regional reporting on the project and its science-fair recognition.
Read coverageAt a glance
Evidence and outcomes
- >97%
- Detection recall
- 0.91
- mAP@0.5
- 99.16%
- System availability
Reported for wildlife approaching the roadway
Object-detection evaluation
Reported during field operation
System design
Perception
- Run YOLOv5 on an NVIDIA Jetson Nano at the roadside.
- Track detections across frames to estimate direction and speed.
Field infrastructure
- Transmit warnings over long-range RF to roadside signage.
- Operate from combined solar and wind power with remote maintenance routines.