In this paper, a novel method for estimating the parameters of multiple chirp signals in additive Gaussian white noise is proposed. The method combines a global optimization theorem with a new Markov Chain Monte Carlo algorithm, called the simulated annealing one-variable-at-a-time random walk Metropolis-Hastings algorithm. It is a computationally modest implementation of maximum likelihood estimation and has no error propagation effect. Simulation results show that the proposed method can give good estimates for the unknown parameters, even when the parameters of the individual chirp signals are closely spaced and the Cram?r-Rao lower bound can be attained even at low signal-to-noise ratio.

In this paper, the frequency scaling algorithm (FSA) has been introduced first to spaceborne spotlight SAR. According to the characteristics that the range time extension of the observed scene is smaller than the width of the transmitted pulse, FSA applied in airborne spotlight SAR dechirps on received echoes and reduces the sampling rate to release the computational burden. However in spaceborne SAR, the range extension is much larger and the bandwidth after dechirping increases instead so that the existing FSA cannot be adopted. To solve this problem, a method to perform directly frequency scaling on chirps in range direction has been proposed. And azimuth compression has also been applied to reduce PRF in azimuth direction. Finally, simulation results of point targets are given to demonstrate the validity and efficiency of the proposed algorithm.

This paper considers the novel application of space-time adaptive processing (STAP) to conformal array radar. Using numerical simulation, we characterize the performance potential of two candidate conformal array designs: a tapered, belly-mounted canoe and a conformal array taking the shape of a chined radome. We find the nonlinear nature of the conformal array design induces clutter angle-Doppler nonstationarity. This nonstationarity leads to covariance matrix estimation errors and a consequent degradation in STAP performance potential. We find these additional losses reside in the range of 4-10 dB for the two array designs under consideration. Finally, we briefly investigate several ameliorating solutions based on localized processing and time-varying weights, achieving performance gains on the order of several decibels to fully mitigating nonstationary behavior over regions of the detection space.

The paper investigates the effect of system geometry of netted sensors on the performance of target position estimation. Each sensor can provide range and bearing measurements, bearing only and range only measurements. Analyses are based on the Cramer-Rao low bound and geometric dilution of precision in 2-dimensional plane. It is concluded that target position estimation accuracy is dependent not only on the range between the sensors and the target but also on the intersection angles between the sensors? line-of-sight to the target. Scenarios in which there are two or three sensors are discussed in detail.

Over-The-Horizon(OTH) HF radars work at heavily congested HF band. It is quite difficult to find broad clear frequency bands for system requirements. The randomly discontinuous spectra (RDS) signal is employed to combat spectrum congestion since it can evade the external interferences in frequency domain. However, the spectra discontinuity of the signal gets rise to high range sidelobes when matching the reflected echo, which is much more difficult for targets detection. So it is indispensable to investigate the technique for sidelobes suppression of the range profile when RDS signal is utilized. In this paper we introduce a new signal processing technique that is radically different from the conventional technique to lower range sidelobes based on suppressing the self-clutter of the radar range ambiguity function (AF) by mismatch filtering. Simulation results show that the peak sidelobe level can be reduced to -30dB while the frequency bands span up to 400KHz.

Printed bow-tie and bow-tie slot antennas are planar-type variations of the biconical antenna that has wideband characteristics. In this paper, we present the effects of adding a tapered metal stub to a bow-tie slot antenna design to enhance the antenna bandwidth for X-band operation. Our study yields a design with 88% bandwidth relative to 10 GHz. The simulation and analysis are performed using the commercial computer software package, Momentum of Agilent Technologies, Advanced Design System (ADS). Verification of the ADS results is performed by using our developed finite difference time domain (FDTD) code and with measurements of the return loss from 7 to 15 GHz.

In this paper a new design for two-frequency MTI radar is introduced. The suggested system can change its frequencies, in each pulse. Therefore, the system is very resistive to electronic war. The analytical results of our calculation show that the system has very high blind speed and in realistic situations, it increases signal to noise ratio, although it widens clutter bandwidth and detects some spurious targets.

In adaptive radar systems, optimal processing for target detection is only possible when the target location in angle and doppler is used to build the processing filter. When this location is not exact, losses occur. These losses translate into a reduction in detection probability. After development of analytical expressions to quantify the effects of this mismatch, two techniques are examined for reducing these effects. These techniques are tested against losses due to doppler mismatch. The first technique involves the use of temporal windows and reduces mismatch losses at the expense of reducing the overall Signal-to-Interference plus Noise Ratio (SINR) of the target. The second technique involves the use of additional filters, and achieves a reduction in mismatch losses without sacrificing maximum SINR. A brief overview of the problems associated with multidimensional (angle and doppler) mismatch is also presented.

This work analyzes the effects of common data windows on STAP algorithms and the non-adaptive signal match processor. The windows are applied both to temporal and spatial dimensions. With the exception of factored approaches, it is shown that STAP algorithm performance decreases when windows are applied. Finally, a Monte Carlo analysis of Probability of Detection is performed on the best windowed/non-windowed combination from each technique evaluated. Because the covariance must be estimated, the results demonstrate that the windowed non-adaptive signal match processor can outperform partially adaptive STAP methods at normalized doppler between 0.25 to 0.75, while partially adaptive STAP algorithms perform significantly better than the windowed signal match processor closer to the clutter normalized doppler.

A vector miss distance estimation algorithm based on source localization is presented in this paper. A small antenna array is used for measuring the movement of a target. Phase differences among antennas and range from the target to the origin can be estimated by some modern spectral analysis method, such as ESPRIT or MUSIC. A closed-form source location estimate is given by the solution of a set of linear equations constructed from the estimates of phase differences and range. Then by utilizing all source locations estimated from different samples, another set of linear equations is formed and unknown vector miss distance parameters can be estimated. Simulation results are provided to demonstrate the effectiveness and feasibility of the proposed method.

Research done in recent years has clearly demonstrated large improvements in clutter suppression and target detection by including elevation adaptivity, otherwise described as 3-Dimensional (3D) STAP. This paper will further quantify the performance gains garnered by 3D STAP by fixing the Degrees Of Freedom (DOF) and varying the array dimensions, to include the equivalently sized linear array. The focus is placed on performance bounds established by matched filter and 3D Cross Spectral Metric (CSM) SINR curves generated with known covariances. The mathematical extension of the CSM from 2D to 3D is shown to be straightforward, thus allowing the CSM to serve as a partially adaptive performance bound for eigenvalue-selection based 3D STAP algorithms.

In this paper, we propose a new detector to test the presence of the radar signal with unknown parameters in unknown variance additive white Gaussian noise. Uniformly Most Powerful Invariant (UMPI) tests and unitary filter banks are combined in a new detector. This problem does not fit the linear model due to the unknown Doppler frequency. It is found that the UMPI test does not exist for such a problem. In our detector, we apply an appropriate linear transformation, which is a unitary filter bank, to the concerned radar signal and the problem is converted to the canonical form. Since UMPI test does exist for the canonical form, then we derive an optimal invariant test for our problem.

A method for evaluating the robustness of constant false alarm rate (CFAR) detectors is presented in this paper, which is based on the powerful methodology of influence function (IF) developed in the literatures on robust statistic. The robustness of different kinds of CFAR detectors can be evaluated and compared by calculating the first derivative of the false alarm probability (FAP) and detection probability (DP) at an underlying distribution, which are named IF-FAP and IF-DP. The comparison of those two IFs among some kinds of CFAR detectors are illustrated. Then it is concluded that the robustness of detector can be asymptotically represented by the IF of clutter power estimator. Finally, according to a robust measure drawn from the IF of clutter power estimator, the "most robust" detectors in three ordered-statistic-based groups are presented.