New tech combines AI and ultrasound to protect wildlife

From the newsletter

Aberdare National Park in Kenya is piloting an artificial intelligence (AI) system to protect eastern black rhino calves from hyena predation. The system utilises machine learning and ultrasound technology to detect and deter hyenas, which are a threat to young rhinos and aims to facilitate the reintroduction of rhinos into the national park. 

• When a hyena is detected, the system emits a high-frequency sound that is inaudible to humans and most animals but repels hyenas.

• The AI-powered ultrasonic deterrent system detects wildlife intrusions in real-time, using object detection and distance sensors. The system deploys adaptive ultrasonic frequencies to deter animals, logs data for analysis, and sends alerts.

More details

• Developed by the Austrian firm IT-Revolutions and led by the Rhino Ark Kenya Charitable Trust in partnership with the Kenya Wildlife Service, this AI system operates by analysing footage from cameras installed along wildlife corridors. It is trained to distinguish hyenas from other species and improves its accuracy over time through self-learning algorithms. This method provides a non-lethal means of predator management, aiming to create a safer environment for rhino calves.

• Currently in its pilot phase, the system is continuously monitored to assess its effectiveness. Researchers are evaluating whether the AI accurately identifies hyenas and whether the ultrasonic deterrent successfully prevents attacks. In addition to technological measures, physical barriers have been implemented to safeguard the rhinos. A 19-kilometre fence has been constructed around the sanctuary, creating an Intensive Protection Zone where rhinos can be closely monitored. The AI detection system is installed at six corridors linking the sanctuary to the broader Aberdare ecosystem.

• The AI-powered ultrasonic deterrent system employs object detection algorithms, distance sensing, and automated actuation to manage human-wildlife conflict. The core technology integrates microcontrollers, servo motors, and ultrasonic emitters. Upon detecting animals near a predefined boundary, the system uses pan-tilt mechanisms for tracking and emits species-sensitive ultrasonic frequencies to discourage further approach.

• Distance sensors continuously assess the proximity of wildlife to the field boundary. When animals breach a certain threshold, ultrasonic waves are activated to initiate non-lethal deterrence. The system's actions are logged into a database for temporal and spatial analysis. Real-time alerts, such as SMS notifications to local authorities, ensure a prompt human response in critical cases.

• The architecture supports modular deployment and is optimised for low-power, remote operations. It enables layered deterrence strategies, with zones triggering increasingly strong responses. Integration with mobile dashboards, IoT monitoring, and predictive analytics allows farmers to manage interventions based on historical wildlife activity, crop conditions, and sensor feedback. The system is designed to adapt to seasonal and environmental variations.

Our take

• Human-wildlife conflict in Africa, driven by habitat loss, climate shifts and expanding human settlements, increasingly threatens both communities and biodiversity. However, emerging technologies provide scalable, non-lethal solutions to foster peaceful coexistence.

• In addition to AI-powered ultrasonic deterrents, other technologies are also addressing this issue. Vodafone’s IoT system detects wildlife and activates deterrents such as flashing lights or bee sounds. Inmarsat enables remote conflict monitoring through mobile satellite technology.

• Beehive fences naturally deter elephants, while GPS collars monitor predator movements to prevent attacks. Local innovations like Lion Lights and GSM motion detectors offer affordable, community-driven protection, encouraging coexistence and reducing the risk of retaliatory killings.