Chair: Richard J. Vaccaro, University of Rhode Island (USA)
C. Theys, G. Alengrin, I3S CNRS/UNSA (FRANCE)
The problem is the detection and the estimation of abrupt changes in a chirp signal. In this case, an exact Generalized Likelihood Ratio (GLR) test cannot be achieved, the system leading to the Maximum likelihood Estimates (MLE) of the parameters being non-linear. The usual solutions are to derive a GLR assuming the signal piece-wise stationary or to supervise estimated parameters of the chirp. We propose two solutions taking into account both slow and fast nonstationarities of the signal. The first consist in a GLR derived from a signal phase model approximation. The second keeps the exact model and uses an approximation of the LR. Delay to the detection is studied and a discussion on estimation of the parameters after change is leaded.
J.P. LeCadre, IRISA/CNRS (FRANCE)
Olivier Tremois, IRISA/CNRS (FRANCE)
Classical bearings-only target motion analysis (TMA) are restricted to constant motion parameters (usually~: position and velocity). However most of the interesting sources have maneuvering abilities, degrading, thus, dramatically the TMA performances. A basic idea consists in modelling the states of the source by an hidden Markov model (HMM for the sequel). This approach is an elegant solution to the maneuvering target tracking problem since it does not require any prior information on the maneuvers, so that its performance does not depend on some accurate criterion that hardly occurs in real scenarios. The main point is then to optimize the observer trajectory using methods derived from the general theory of dynamic programming. A main difficulty arises from the partial observation of the source state. This leads us to model the problem by means of POMDP (Partially Observable Markov Decision Process) for which efficient solutions exist.
Constantino Rago, University of Connecticut
Peter Willett, University of Connecticut
Roy Streit, Naval Undersea Warfare Center (USA)
Here we analyze the tracking characteristics of a new data- association/tracking algorithm proposed by Streit and Luginbuhl, the Probabilistic Multi Hypothesis Tracker (PMHT). The algorithm uses a recursive method (known amongst statisticians as the Expectation-Maximization or EM method) to compute in an optimal way the associations between measurements and targets. Until now, no comparative performance analysis has been done. In this paper, we compare the performance of this new scheme to that of a commonly used tracking algorithm, the Joint Probabilistic Data Association Filter (JPDAF).
P. Gounon, CEPHAG, ENSIEG (FRANCE)
S. Bozinoski, CEPHAG, ENSIEG (FRANCE)
We present in this paper a high resolution method for the joint estimation of Directions Of Arrival (DOA) and Travel Time (TT). This algorithm applies to active antenna for wich the transmitted signal of some sources is known. We show how to take into account the a priori information about signal in MUSIC. The method is presented and a non asymptotic statistical performance analysis using perturbation expansions is applied. The major result is a formula for mean-squared error of the DOA and TT estimations. Simulation results verify the analytically predicted performance.
Wayne T. Padgett, Rose-Hulman Institute of Technology (USA)
The detection of nonstationary random signals is an important sonar problem which also has potential applications in diverse areas such as biomedical signal processing and spread spectrum communications. The primary problem with applying a powerful test like the generalized likelihood ratio test (GLRT) is the computational effort required to search for the maximum likelihood model parameters for the observed signal. The computation required is multiplied many times over when a signal parameter is nonstationary. A computationally efficient detector of nonstationary Gaussian random signals based on the GLRT was presented at ICASSP94. A slightly enhanced version of the detector is described below, along with new simulation results demonstrating that the detector performs nearly optimally and is quite robust to signal model inaccuracy.