<title>Kinematic ranging for IRSTs</title>

Abstract

A 3-dimensional kinematic ranging algorithm for IRSTs is described, capable of real-time operation at greater than 25 Hz. The algorithm implements two 2-dimensional extended Kalman filters to estimate range in azimuth and elevation which are then combined to give the 3-dimensional range. Also included are algorithms for target manoeuvre detection and for determining the confidence of the range estimate.Simulation results for the algorithm are presented, which show that accuracies of better than 20 % can be achieved at long range, depending on the measurement accuracy.

Acoustic SLAM

“…1) The number of sources and their states relative to the observer as in (5) and (6) are unknown and time-varying. 2) The model in (4) and (10), mapping from the 2D DoAs to the 3D states, presents an underdetermined system. 3) The non-linear wrapping in (10) results in a non-Gaussian Probability Density Function (pdf) of the DoA estimates.…”

“…The state of each component in (29) can be predicted at t using the source dynamical model in (4). The predictions are subsequently updated by inferring information from the DoA estimates in (10). As derived in Appendix A, the predicted and detection PHDs in (19) and 20thus are:…”

“…The observer reports are generated from (3). DoA estimates are generated from (10) with λ c = 0 and p d = 1. The covariance terms, σ 2 w ,γ t , σ 2 w ,v t , and R t,m are assumed known.…”

“…For moving observers, spatio-temporal diversity of the sensors can be exploited for kinematic ranging of the 3D source positions from the DoA estimates [10], [11]. However, the DoA estimates are specified relative to the observer.…”