SWARM-IS: Adaptive sensor swarm search strategy and simulation framework for predator surveillance and control

Introduced mammalian predators are causing declines in many of New Zealand's unique, native species. The New Zealand government has an ambitious Predator Free 2050 goal that aims to eradicate seven predator species by 2050. An important part of monitoring progress towards this goal is the ability to detect the presence of predators in different environments. Existing research does not take advantage of recently developed AI-enabled sensor nodes. In this paper, we present a novel adaptive sensor swarm search strategy, called SWARM-IS, that determines where in the landscape predator-detecting sensor nodes should be placed based on the detection histories reported by the sensor nodes themselves. Additionally, we develop a simulation framework to evaluate our new strategy's performance compared to the Systematic and Random search strategies. With our simulation framework, we perform two experiments varying the ecological conditions and detection efficacy. We show through these experiments that our swarm-based search strategy is able to detect up to 20% more predators in a given area compared to Systematic and Random search strategies.