Human Elephant Conflict (HEC) is an increasing problem in elephant habitat countries, including Zambia. It causes immense tension in villages, damage to crops, stores and homes.
Game Rangers International aims to address shortcomings in previous HEC mitigation methods by using ground-breaking technology created by Hack the Planet, a Dutch team, who develop meaningful technology for conservation. This innovation addresses known challenges to existing repellent methods, namely: high cost, high level of stress imposed on the elephants, and low long-term effectiveness (due to habituation where elephants grow accustomed to the deterrent).
This action-orientated pilot aims to use ‘smart’ technology to safely deter elephants by using a combination of a smart camera trap and a light and sound alarm to be installed near key conflict areas. The Artificial Intelligence (AI) smart camera element monitors approaching elephants and triggers an ‘at-a-distance’ alarm to deter the elephants. The Repeller will start producing sounds (both natural and human made, including songs) with flashes of light. Given that elephants are incredibly intelligent, producing one sound the whole time will not work, as they will eventually learn that it is something automatically playing and dismiss it. Therefore, the repeller has a large collection of audio files which are played at random to ensure that the elephants never hear the same sound twice. Since the camera trap and repeller are at least 500 meters apart, the elephants will not be startled but will gradually hear more noises coming from the distance. The distance element serves as an additional barrier, reducing elephant likelihood to grow accustomed to the alarm, and increasing community safety. The idea behind this is that it will simulate the natural activity of people and make the elephants wary to approach. The camera trap collects and stores valuable information about the incident (i.e., time of approach, location) which can be useful to trigger the rapid deployment of response teams.
A key strength of this project is that we will address the issue of ‘research fatigue’ which community members often face when offered solutions that don’t work. Due to the unique position that GRI have with access to orphan elephants (and the belief that where possible they should be contributing to wider conservation outcomes) we can initially pilot appropriate methods on them and get video evidence of the system working to give the farmers increased confidence in the method by being able to ‘see’ it for themselves.
Therefore, before deploying the idea into the communities, we are conducting two trials: the first, around the GRI-Kafue Release Facility which will test the repellers on the released elephants and the second, in the release area vicinity which will target wild elephants. The main objective of the trials is to assess the effectiveness of the elephant repellers on African elephants, with an additional camera trap to gather more detailed behavioural responses allowing greater understanding of the impact of the different sounds, enabling the system to be further improved.
Since the deployment of the repellers, we have received alerts of elephants being captured by the smart cameras via Earth Ranger. Earth Ranger is a real-time software solution for park managers and wildlife researchers which collects, integrates and displays historical and real-time data. For this project the AI picks up seeing an elephant or other animal and feeds that information to Earth Ranger which tells us on a map if an “elephant” or “other” (any other animal) has been detected.
The sound box near camp has been triggered several times by both the released and wild elephants. We have so far heard some interesting songs such as Celine Dion, Bach and even some Dutch songs! Chamilandu and the released herd who were passing near camp one afternoon, were so startled by the music coming from the sound box (which had been triggered by wild elephants) that they ran towards camp, but instead of a reassuring welcome home they were chased out by a keeper, since we do not want to encourage them to return.
From the trials conducted so far, we have gained very encouraging results, already capturing video and GPS collar support to evidence that the system is effective. The tests have already enabled Hack the Planet to improve the system, so we are very enthusiastic to move the project onto the Community Phase, aiming to prove the concept right at the front line, gather relevant data, allowing for additional development to improve effectiveness and lower the cost to make it readily available for those desperately in need.