L. Timmoneri

Knowledge-based system for multi-target tracking in a littoral environment

The paper addresses how to efficiently exploit the knowledge-base (KB), e.g. environmental maps and characteristics of the targets, in order to gain improved performance in the tracking of multiple targets via measurements provided by a ship-borne radar operating in a littoral environment. In this scenario, the nonhomogeneity of the surveillance region makes the conventional tracking systems (not using the KB) very sensitive to false alarms and/or missed detections. It is demonstrated that an effective use of the KB can be exploited at various levels of the tracking algorithms so as to significantly reduce the number of false alarms, missed detections, and false tracks and improve true target track life. The KB is exploited at two different levels. First, some key parameters of the tracking system are made dependent upon the track location, e.g., sea, land, coast, meteo zones (i.e., zones affected by meteorological phenomena) etc. Second, modifications are introduced to cope with a priori identified regions nit hi high clutter density (e.g. littoral areas, roads, meteo zones etc.). To evaluate the behavior of the proposed knowledge-based tracking systems, extensive results are presented using both simulated and real radar data

Comparison of recursive and batch processing for impact point prediction of ballistic targets

The paper deals with the problem of impact point prediction of ballistic targets (BT) by processing measurements acquired either by 3D surveillance or multifunctional phased-array radars. It is assumed that the radar acquires a limited number of measurements that do not encompass the whole target trajectory; thus the established target track has to be extrapolated ahead in time in order to predict the coordinates of the impact point. In this paper we compare performance of batch (i.e. maximum likelihood estimation, MLE) and recursive (extended Kalman filter EKF and unscented Kalman filter UKF) filtering techniques.

Use of a graphics processing unit for passive radar signal and data processing

The article discussed two main topics: 1. It illustrated the need for increasing the performance of the passive sensor signal processor. The GPU appears to be a solution especially for the computational request of a dual-band system simultaneously receiving and processing FM and DVB-T data. 2. It presented of the AULOS passive system already in the field.

CRLB and ML for parametric estimate: new results

In the present paper, we calculate the Cramer-Rao Lower Bound (CRLB) for the Pd 0 case, accounting both for missed detections and false alarms, for an estimation problem in a surveillance system where the measurements are acquired by a radar. The analysis, which is limited to the parameter estimation for a deterministic steady-state target motion, is a further innovation in the computation of the CRLB of maximum likelihood estimators for deterministic steady-state models; in fact, the classical theory applies only in the unrealistic case of Pd = 1 and Pfa = 0. The results obtained by means of a Monte Carlo simulation successfully validate our extended enhanced estimation method.

A Real Time Test Bed for 2D and 3D Multi-Radar Tracking and Data Fusion with Application to Border Control

This paper describes part of the activities required to model and assess the performance of a national integrated system (NIS) in the context of homeland security and in particular on border control issues. The border control requires an accurate surveillance of the terrestrial, maritime and overland boundaries, thus involving a wide number of heterogeneous sensors, command and control centres, platforms and communication networks. One of the layers of the integrated system is composed by radar sensors; these sensors can be ground or ship based, 1D or 3D. The paper deals with the problem of correlation and fusion of track data pertaining to ground and ship based, 2D and 3D radar sensors directly in the radar sites. Then, the improved (i.e. more accurate) information can be soon made available for a number of radar functions, such as: (i) re-pointing of the beam along the threat direction of arrival, (ii) energy and time management, (iii) cue of other sensors (i.e. infrared) or proper reaction means (patrol boat), (iv) preliminary classification of potential hostile targets. The updating and testing of the data extractor (DE) of a notional surveillance radar system is presented; the modified hardware and software is capable of acquiring, managing and fusing tracks pertaining to the radar system housing the DE and to other systems connected to the DE itself

Classification and launch-impact point prediction of ballistic target via multiple model maximum likelihood estimator (MM-MLE)

The paper deals with the problems of (i) launch and impact point prediction (LPP, IPP) of ballistic targets (BT) and (ii) BT classification by processing measurements acquired either by 3D surveillance or multifunctional phased-array radars. It is assumed that the radar acquires a limited number of measurements (plots) that do not encompass the whole target trajectory; thus, the established target track has to be extrapolated ahead in time in order to predict the coordinates of the impact point. A procedure based on multiple model maximum likelihood estimator (MM-MLE) has been conceived and tested using a Monte Carlo simulation approach; the parameters selected for testing are the probability of BT correct classification (Pcc), the IPP and the LPP. The new procedure is compared with the estimator described in A. Farina et al. (2004).

Challenging Issues in Multichannel Radar Array Processing

This paper is focused on multichannel radar array processing and addresses some challenging issues concerning: the adaptation in nonstationary and nonhomogeneous environments, the value of Space-Time Adaptive Processing (STAP) in an operational High Frequency (HF) radar system, the benefits resulting from the joint use of multiple (not colocated) transmitters and receivers, and the parallel implementation of the multichannel processor. The work follows on where the importance of a multichannel processing is demonstrated with reference to airborne, ground-based, and Over The Horizon (OTH) radar systems.

Tracking a ballistic target by multiple model approach

Radar tracking of a ballistic object flying in the Earths atmosphere is a very complex issue to cope with, due to the need of (suboptimal) nonlinear filtering techniques. When the characteristics of the target are poorly known, and an identification problem is added, a solution is represented by a multiple model approach. This paper investigates the problem by evaluating a number of parameters which affect the solution. The multi modal approach is compared with a generic extended Kalman filter. A theoretical limit for the performance is computed by means of the posterior Cramér-Rao lower bound.

Adaptive spatial processing applied to a prototype passive covert radar: Test with real data

This paper describes an upgrade of the design of a passive covert radar (PCR) [1] to add the capability of adaptively cancel the co-channel interference in the spatial domain. The PCR exploits a single non co-operative frequency modulated (FM) commercial radio station as its transmitter of opportunity. The system has been conceived in order to detect low flying and low radar cross section (RCS) targets even in disturbed and hostile electromagnetic environments. In the first phase of the PCR design, the co-channel interference has been removed via an adaptive algorithm operating in the time domain. In the second phase of the PCR development, it has been decided to improve the adaptive cancellation performance operating in two domains separately; first the adaptive spatial cancellation is applied to reduce the co-channel power which is definitely removed in a second stage via an adaptive temporal processing. Future step will be the contemporaneously process of data in a 2D domain via a STAP (space time adaptive processing) algorithm.

CRLB with pd<1 for bearings only fused tracks

Abslracl - This paper presents mathematical procedures to calculate the Cramer-Reo Lower Bound (CRLB) for a target track obtained byfusing two tracks provided by two sensors that build up their tracks on the basis of bearings only measurements. It is assumed that the two sensors could have a detection probabilities Pd less than 1. The methods can be easily extended to the fusion of more than two sensor tracks.