Adaptive Aerial Localization Using Lissajous Search Patterns

This work presents an adaptive approach to cooperative aerial search and localization (SAL) which implements Lissajous search patterns and non-Gaussian observation likelihoods to preserve high target information. The adaptive component of the framework utilizes a simultaneous estimation and modeling technique to both estimate agent states and correct their motion models. In order to maximize the information available about a target even when it is not observed by a search agent, multi-Gaussian observation likelihoods are continuously generated for each agent and then fused across the search team. Monte Carlo simulation studies show that the proposed adaptive localization framework outperforms standard filtering techniques by significant margins, for a wide range of parameter values. The differential entropies of fused target likelihoods are studied for various multiagent Lissajous pattern configurations, leading to the derivation of optimal Lissajous parameters for cooperative SAL. This work has relevance for SAL applications in rescue, safety, and defense sectors, offering a robust solution to target localization when
a priori
target motion information is unavailable.