Dr SP le Roux’s research combines machine learning and digital signal processing to help park rangers and behavioural ecologists keep an eye on the magnificent beasts. “I’ve always had a passion for nature conservation and technology,” said Le Roux.
During his electrical engineering doctorate programme at the university, Le Roux designed a biotelemetry tag that attaches to the rhino and detects its movements. Using GPS tracking and machine learning algorithms, the device will keep rangers informed.
For a certain period of time the tag will record all the rhinos’ movements. Conservationists will review the data and flag specific situations. For instance, if a rhino is lying down too much or running away from danger.
In future, when the device encounters similar situations, it will alert park rangers who can then send drones or go on foot to see what is happening.
The goal is for the device to successfully alert rangers during a poaching situation. The technology would flag situations in which a rhino is running at a strange time of day or location. Because it uses real-time communication, rangers would potentially be able to intercept and apprehend poachers.
The tags have successfully been put to trial on a flock of sheep and on rhinos at a private game reserve. The sheep proved that the device could be useful by alerting farmers if an ewe is sick. By using the data to see which ewes were lagging behind or lying down too much, farmers were able to easily pick sick ewes out of the group for treatment.
The device’s potential is far-reaching. Besides alerting rangers about looming threats, it will also give behavioural ecologists insights into species behaviour patterns that have never been so closely monitored. The device might also be useful for farmers monitoring livestock.
The testing phase of the collar-tag has been completed. In the future Le Roux hopes to see the device deployed on a larger scale. Existing biotelemetry tags generally do not provide real-time data and can only monitor the animals for a short period.
Le Roux’s design delivers information about an animal every 6.5 seconds and is able to monitor movements and patterns for long periods of time.@TheCapeArgus