A prologue to estimating the intent of a potential rhino poacher

Abstract

A dramatic increase in rhino poaching, if allowed to continue at current rates, is threatening South Africa's rhino population with extinction. The available detection and prevention systems are not sufficient in solving this crisis. Targets are detected, but their intent is unknown, thus a system is proposed that fuses the available data and infers a potential poacher's intent. A mixture of behaviour modelling and intent estimation, coupled with a suitable computational intelligence technique will address the problem of how to infer whether a person is a rhino poacher. This article sheds some light on the current systems that are in place and we offer a formulation of the rhino poacher problem in the Kruger National Park (KNP). The proposed research is to model the terrain of the KNP, as well as the behaviour of rhinos and humans. A simple generic and high-level model will initially be used and the complexity adjusted as the project progresses. The model will be based on experience and observations of game rangers, park officials, South African Police Service (SAPS), South African National Defence Force (SANDF), and other groups that see to the safety of the animals of the KNP. Data will be generated with the model and used to train a chosen computational intelligence algorithm. The contribution will be to show whether computational intelligence techniques can be effective in estimating the intent of potential rhino poachers in a statistical model of rhino poacher behaviour in the KNP. With the proper infrastructure this work could be used to provide decision support to park officials. This is extremely important for the rapidly waning rhino population in South Africa, as well as the rest of the world.