Compressive Detection of Random Subspace Signals

Abstract

The problem of compressive detection of random subspace signals is studied. We consider signals modeled as $\mathbf{s} = \mathbf{H} \mathbf{x}$ where $\mathbf{H}$ is an $N \times K$ matrix with $K \le N$ and $\mathbf{x} \sim \mathcal{N}(\mathbf{0}_{K,1},\sigma_x^2 \mathbf{I}_K)$. We say that signal $\mathbf{s}$ lies in or leans toward a subspace if the largest eigenvalue of $\mathbf{H} \mathbf{H}^T$ is strictly greater than its smallest eigenvalue. We first design a measurement matrix $\mathbf{\Phi}=[\mathbf{\Phi}_s^T,\mathbf{\Phi}_o^T]^T$ comprising of two sub-matrices $\mathbf{\Phi}_s$ and $\mathbf{\Phi}_o$ where $\mathbf{\Phi}_s$ projects the signals to the strongest left-singular vectors, i.e., the left-singular vectors corresponding to the largest singular values, of subspace matrix $\mathbf{H}$ and $\mathbf{\Phi}_o$ projects it to the weakest left-singular vectors. We then propose two detectors which work based on the difference in energies of the samples measured by two sub-matrices $\mathbf{\Phi}_s$ and $\mathbf{\Phi}_o$ and prove their optimality. Simplified versions of the proposed detectors for the case when the variance of noise is known are also provided. Furthermore, we study the performance of the detector when measurements are imprecise and show how imprecision can be compensated by employing more measurement devices. The problem is then re-formulated for the case when the signal lies in the union of a finite number of linear subspaces instead of a single linear subspace. Finally, we study the performance of the proposed methods by simulation examples.Comment: 33 pages, 11 figures, Revised versio

Jointly Optimized Target Detection and Tracking Using Compressive Samples

“…In [12], a random signal with a known sparsity can be detected by a compressive detector with a fixed subspace. However, this detector is not suitable for time-varying sparse signals.…”

“…In [10] and [11], compressive Bayesian detectors are proposed under the condition that the prior probabilities of two hypotheses are available. In addition, a random subspace detector is studied for unknown-parameter sparse signal detection [12], but the fixed subspace must be acknowledged before detection, which may limit the application of this detector. The third category consists of detectors designed by optimizing the transmitted waveform.…”

“…First, CSP-JDT utilizes this prior information to construct the measurement matrix and determines the subspace where the target may exist. Different from the fixed subspace in [12], the proposed subspace is adaptively changed with the prior information, and it can further reduce the calculation amount by reducing the detection range. Then, CSP-JDT uses the prior information to jointly locate and track the target in the subspace range, which can achieve a global optimization.…”

Outage Probability Minimization for Energy Harvesting Cognitive Radio Sensor Networks

“…The allocation of time durations for the two phases is illustrated in Figure 1 , where the channel acquisition phase and the transmission phase consume and fractions of one time slot, respectively, and is referred to as spectrum sensing overhead for time slot t , which can be altered to optimize the performance of the system. For the channel acquisition phase, the SU senses the status of the spectrum with time through the energy detection technique [ 38 ]. As the complexity is roughly linear in sensing duration, we can assume that the energy consumption for sensing is proportional to with a constant sensing power [ 32 ], namely: …”

Tree Copula Theory Based Fusion And Compressive Sensing For Activity Detection using Multi-Modal Data

“…In case of multi-sensor recording, when the number of sensors is greater than that of the number of sources termed as overdetermined case, then the recovery of the original signal can be done using the linear combination of obtained mixed signals [8]. In order to model the complex dependencies over multivariate data for the evaluation of joint pdf, copula theory is used in [2][3][4][5][6][7][8]. Hence, the advantages of using the copula theory for LR based detection in the presence of multi modal data requires more computational price [5].…”