Find the hoofbeasts, tell the drivers, save the world.
This is a mirror of our engineering capstone project, sponsored by Garmin.
Wildlife entering roads is a major hazard for drivers. Environmental obstructions, weather conditions, time of day, and driver fatigue contribute to the difficulty drivers have distinguishing approaching wildlife. This problem is traditionally solved with fences and wildlife overpasses, but these are expensive and have major impacts on wildlife mobility.
Our solution, a wildlife detection system, is installed as a non-obstructive, wireless grid of sensors near the road. The sensor network detects nearby wildlife and alerts drivers of the approaching danger with LED signals.
Each node in our system is fitted with two passive infrared or infrared laser sensors, providing redundant detection coverage of the network. We trained the system’s behavior using a genetic machine learning algorithm to optimize performance. Garmin’s ANT BLAZE mesh network technology makes the system resilient to failure of individual nodes, ensuring that warnings reliably reach drivers. Network configuration and status monitoring are enabled by our PC app, which interfaces wirelessly with the sensor nodes over ANT.
By giving drivers advance notice of wildlife threats, our system has the potential to reduce vehicle collisions with wildlife without the expense and environmental impact of existing solutions.