Search

Micro-Doppler effect due to target micro-motions induce additional frequency components in the received radar signal. In addition to other target parameters, these micro-Doppler signatures can be used for target identification. For targets with high frequency micro-motion we observe that the conventional extraction of micro-Doppler signature in Frequency Modulated Continuous Wave (FMCW) radar fails due to insufficient sampling frequency along the slow-time. In this paper, we first develop a signal model for FMCW radar in the presence of a target with micro-motions. It can be inferred from the model that the micro-Doppler signature is preserved in the received signal. We then propose an approach to extract high frequency micro-Doppler signatures for FMCW radar, and verify it using simulation studies and experimental results.

Besides the detection of targets and estimation of their parameters, micro-motions of various target parts, that give rise to micro-Doppler signatures, are extracted from Radar acquisitions for target identification. Frequency modulated continuous-wave (FMCW) radar is identified to have the potential to perform both the activities. Due to high rotation rate of the rotor blades in unmanned aerial vehicles (UAV), the micro-Doppler frequency variation is significantly high giving rise to aliasing effect. In this work, we first study the signal model of FMCW radar in the presence of UAV. A new approach is then presented for the study of high frequency micro-Doppler signatures. Simulation example and experimental data show the efficacy of the approach for micro-Doppler signature analysis.

This paper proposes the generalized MUltiple SIgnal Classification (MUSIC)-like algorithm for robust MUSIC-like processing for underwater applications. The solution proposed in this paper is to generalize the noise correlation assumption and include a noise correlation model in its problem formulation. By doing so, the proposed generalized MUSIC-like algorithm is able to provide robust MUSIC-like performances in any noise condition, so long as the noise correlation property of the environment is known partially. Results from simulations and real data processing show that our proposed algorithm is able to suppress spurious peaks caused by mismatched noise assumptions in standard MUSIC-like algorithms. The bound of the controlling parameter denoted by $\beta $ for the proposed generalized MUSIC-like algorithm is also discussed in this paper. Performance study using Monte Carlo simulations shows that the proposed generalized MUSIC-like algorithm has the same resolving power as the MUSIC method but slightly poorer accuracy in direction-of-arrival (DOA) estimation. This paper also presents the results from real data processing by the generalized MUSIC-like algorithm and demonstrates better resolving power than the Capon and MUSIC algorithms used consistently in the experiment.

When the number of sources is inaccurately estimated, it is well-known that the conventional subspace-based super-resolution direction-of-arrival (DOA) estimation techniques provide inconsistent spatial spectrum, and hence the DOA estimates. In this work, we present a novel technique which provides resolution capability comparable with that of the super-resolution techniques. While the working principle of the proposed technique is similar to that of the minimum-norm algorithm, the algorithm is insensitive to the estimated model-order. Simulation studies show that the proposed technique is advantageous over the use of subspace-based techniques with the number of sources estimated by well-known model order estimation techniques.

With the bandwidth of speech signals extending over several octaves, the spatial Nyquist criterion constrains the microphone array design. Violating this criterion by increasing microphone spacing in order to achieve high resolution introduces ambiguity in identifying the source directions due to the aliasing components. In this work, we investigate the effect of spatial aliasing on the direction-of-arrival (DOA) spectrum due to wideband sources. Noting that the extent of aliasing is frequency dependent, we propose a multi-stage scheme for speech DOA estimation following a subband decomposition. To observe the advantage of this scheme, we verify it with the steered minimum variance distortionless response (STMV) and approximate kernel density estimators. The performance is evaluated with simulations and recorded room impulse responses.

In this paper, we propose a new technique to estimate wideband source directions from the sensor snapshots without requiring to know the number of sources present in the scenario. This work is motivated by the fact that the existing model order estimation (number of sources) techniques for wideband source scenario are either inaccurate or computationally expensive. Direction-of-arrival (DOA) estimation is realized using a beamformer framework which imposes nulls in the spatial spectrum along the source directions. The null width along the frequency axis is widened by introducing a new data dependent term into the optimization problem, thus achieving wideband capability. Furthermore, the temporal processing of the data snapshots drastically reduces the number of snapshots required for wideband DOA estimation. The effectiveness of the proposed formulation is studied with simulated experiments.

Various array processing techniques applied to uniform linear arrays are involuntarily realized using structures that are analogous to finite impulse response filters. This observation leads to the following question: “can we extend infinite impulse response (IIR) filtering to array processing?”. In this paper, we introduce the concept of IIR array in spatial domain. Note that IIR array here does not mean time-domain IIR filtering for array beamforming which is commonly understood. This paper is dedicated to the study of an alternate approach for array signal processing which defines IIR structure in spatial domain. To illustrate the applicability of the concept of IIR array, we propose a new direction-of-arrival estimation technique as well as a beamformer with the spatial domain IIR array implementation. The performance of the proposed methods are comparable to the existing techniques. These illustrations are intended to introduce a new approach which potentially can offermore degrees of freedom to control the performance of the array and reduce the complexity of the system for a desired performance. Accepted version

Due to practical considerations the microphone spacing is increased to achieve improved resolution by violating the spatial Nyquist criterion. Accompanied aliasing components adversely affect the identifiability of the source direction peaks. We investigate the effect of aliasing on the spatial spectrum of the steered minimum variance distortionless response (STMV) method and propose a novel multi-stage scheme assisted by subband decomposition for suppressing aliasing components. The performance of the proposed technique, evaluated with simulations and recorded room responses, reflects the improvement in the identifiability of accurate source directions under aliasing conditions.

The problem of direction-of-arrival (DOA) estimation in dispersive medium is considered in this work. For accurate localization of the sources in seismic and underwater acoustic applications, speed profile is estimated and employed. In contrast to this approach, we propose a new algorithm for wideband DOA estimation with unknown speed profile across the source bandwidth. This insensitivity to dispersion is accomplished by incorporating soft derivative constraint into the optimization problem. The presented simulations exhibit the effectiveness of the proposed algorithm in estimating the source directions in dispersive medium.

Array processing has been an active research area for several decades. The advent of new signal processing techniques has maintained this topic afresh in the research community with new challenging problems. The estimation of direction-of-arrival (DOA) for instance has evolved from high-resolution to superresolution techniques, and of-late, is heading towards increasing the available degrees of freedom. Excited by these developments and the awaiting potential applications, we have considered the study of DOA estimation for narrowband and wideband sources in this thesis under various conditions. Within an array processing system, the inter-dependence between the model-order estimation, DOA estimation and beamforming tasks reflects the sensitivity of one task to the outcome of the other. In view of this, beamformers have been designed in the past to incorporate robustness against look direction mismatch and array manifold errors. However, existing DOA estimation techniques are sensitive to the accuracy of estimated number of sources. In order to overcome this limitation, we propose a new narrowband DOA estimation technique which substitutes the noise subspace eigenvectors with a weight vector matrix. This allows one to obtain the spatial spectrum with unknown number of sources. Any error in model order estimation will therefore have no impact on the accuracy of DOA estimates. Estimating the number of sources in the presence of wideband sources is a very challenging task considering the fact that existing techniques retrieve the model order either from a coherently-averaged covariance matrix or by the maximum likelihood approach. While the estimated model-order from the former method is susceptible to the initial estimates, the latter technique is computationally expensive. We therefore present a time-domain DOA estimation technique which provides distinct peaks along the source directions in its spatial spectrum without estimating the number of sources. The underlying idea relies on the array manifold approximation using Taylor series expansion across the signal bandwidth. The undesired derivative components are then suppressed by the proposed optimization problem. The effectiveness of this technique is verified with a detailed mathematical analysis and simulations. With finite-ordered Taylor series expansion, the array manifold approximation is accurate for sources with a percentage bandwidth less than 30%. For larger source bandwidth, the estimation accuracy of the time-domain technique decreases. We therefore transform the problem to frequency domain and perform DOA estimation on a regulated signal bandwidth. Environmental factors such as multipath, dispersion and scattering adversely affect the performance of existing DOA estimation techniques in many applications such as radio wave communication, seismic and underwater acoustic applications. Existing techniques such as matched-field processing incorporate the speed profile and introduce robustness to random perturbations in speed. However, the estimation of speed profile is itself challenging and inaccurate many times. We therefore consider redefining array manifold approximation such that robustness can be incorporated to dispersion. The optimization problem introduces derivative compensation with respect to the wavenumber which absorbs the effect of dispersion in the signal model. With this approach, one requires to only estimate the propagation speed at only a reference frequency instead of the entire source bandwidth. DOCTOR OF PHILOSOPHY (EEE)

In this paper, we present a new approach for wideband direction of arrival (DOA) estimation incorporating derivative constraints into the optimization problem. The Taylor series expansion of steering vector with respect to frequency exhibits the dependence of sensor snapshots on the derivatives of the array manifold matrix, suppression of which reduces the data to follow the narrowband data model. Two new formulations following a beamformer framework are developed based on this observation. While the first formulation estimates accurate source directions with apriori knowledge of the signal subspace dimension, the second formulation provides DOA estimates without estimating the number of sources. The coherent averaging at the reference frequency enables the two formulations to handle coherent sources. The effectiveness of the proposed formulations is illustrated with simulations.

We address the problem of human footstep detection using data recorded by a single tri-axial geophone. It is observed that footstep signature recorded using a vector-sensor is characterized by signal polarization, which, when exploited effectively, has the capability to identify footsteps at increasing source-sensor distances compared to existing techniques. We quantify the effect of signal polarization by fitting a great-arc using spherical linear interpolation (SLERP) to the data vectors after normalization. Furthermore, the signal polarization metric, which provides extended detection range, is combined with signal energy to form a robust polarization-cum-energy metric for efficient detection. Experimental results are presented to substantiate the performance of this technique.

In this paper, we consider the problem of underdetermined blind source separation for anechoic speech recordings. Existing two-stage methods which first estimate mixing matrix and then separate sources are suitable only for instantaneous mixtures and do not cater for anechoic speech recordings. Other time-frequency (TF) methods based on binary masks are found to have limited performance. We here propose a new two-stage technique which includes frequency normalization to estimate mixing matrix followed by a source separation stage involving denormalization process to estimate frequency dependent mixing matrices. Experimental results are provided to demonstrate the advantage of the proposed method over other methods.

In this paper we propose a new footstep detection technique for data acquired using a triaxial geophone. The idea evolves from the investigation of geophone transduction principle. The technique exploits the randomness of neighbouring data vectors observed when the footstep is absent. We extend the same principle for triaxial signal denoising. Effectiveness of the proposed technique for transient detection and denoising are presented for real seismic data collected using a triaxial geophone.