# Cubical and spherical directional array-based particle source detection with poisson statistics

## TL;DR

This paper compares different statistical tests for detecting particle sources using cubical and spherical arrays, showing that the truncated mean test performs best.

## Contribution

The paper introduces a comparative analysis of detection tests for dynamic particle sources using directional arrays and demonstrates the superiority of the truncated mean test.

## Key findings

- The truncated mean test (TMT) outperforms existing tests like the mean difference test and GLRT in detection performance.
- The analysis includes scenarios with both stationary and moving targets for cubical and spherical arrays.
- Probabilities of detection and false-alarm are evaluated for different observation types and array geometries.

## Abstract

This paper investigates detecting a far-field dynamic particle source using cubical and spherical directional arrays. First, we present a discussion on likelihood ratio test (LRT), generalized LRT (GLRT) and truncated mean test (TMT) in case of a cubical array, considering scenarios with stationary and moving targets. Further, two underlying sub-cases are considered, including the sets of correlated and independent and identically distributed observations. Next, we consider the case of a spherical array where LRT, GLRT and TMT are discussed. In all cases, we present an analysis in terms of probabilities of detection and false-alarm for the associated hypothesis testing problems. Through simulations, we show that TMT outperforms the other existing tests in the literature, namely the mean difference test, the source intensity test, and the GLRT.