Showing papers by "Alexander M. Haimovich published in 2015"

Beamforming Design for Joint Localization and Data Transmission in Distributed Antenna System

Abstract: A distributed antenna system whose goal is to provide data communication and positioning functionalities to mobile stations (MSs) is studied. Each MS receives data from a number of base stations (BSs) and uses the received signal not only to extract the information but to determine its location as well. This is done based on time-of-arrival or time-difference-of-arrival measurements, depending on the assumed synchronization conditions. The problem of minimizing the overall power expenditure of the BSs under data throughput and localization accuracy requirements is formulated with respect to the beamforming vectors used at the BSs. The analysis covers both frequency-flat and frequency-selective channels and accounts for robustness constraints in the presence of parameter uncertainty as well. The proposed algorithmic solutions are based on rank-relaxation and difference-of-convex programming.

Non-Coherent Direction of Arrival Estimation from Magnitude-Only Measurements

Abstract: We consider the classical Direction of arrival (DOA) estimation problem in the presence of random sensor phase errors are present at each sensor. To eliminate the effect of these phase errors, we propose a DOA recovery technique that relies only on magnitude measurements. This approach is inspired by phase retrieval for applications in other fields. Ambiguities typically associated with phase retrieval methods are resolved by introducing reference targets with known DOA. The DOA estimation problem is formulated as a nonlinear optimization in a sparse framework, and is solved by the recently proposed GESPAR algorithm modified to accommodate multiple snapshots. Numerical results demonstrate good DOA estimation performance. For example, the probability of error in locating a single target within 2 degrees is less than 0.1 for ${\rm SNR} \geq 15~\hbox{dB}$ and one snapshot, and negligible for ${\rm SNR} \geq 10~\hbox{dB}$ and five snapshots.

Ziv–Zakai Bound for Joint Parameter Estimation in MIMO Radar Systems

Abstract: Local bounds, such as the Cramer–Rao bound (CRB), provide inaccurate predictions under low signal-to-noise ratio (SNR) conditions. Global bounds are capable of providing more accurate predictions of the performance of estimators over the full range of SNR. In this paper, we derive the Ziv–Zakai bound (ZZB) for joint location and velocity estimation of a target illuminated by a non-coherent multiple-input multiple-output (MIMO) radar employing orthogonal waveforms and widely spaced antennas. The setup captures the multistatic nature of the target gains, where each pair of transmit-receive elements experience independent gains governed by the Swerling type 1 model. The target returns are observed in the presence of spatially and temporally independent Gaussian clutter-plus-noise. The ZZB for joint delay and Doppler estimation for single-input and single-output (SISO) radar is also developed. We show that the ZZB is a comprehensive metric that captures the effect of the SNR, the ambiguity function (AF) and other parameters of the radar systems. The effects of different system configurations are explored through numerical studies. The results are useful for the analysis of both active and passive radars.

Cloud Radio-Multistatic Radar: Joint Optimization of Code Vector and Backhaul Quantization

Abstract: A multistatic radar set-up is considered in which distributed receive antennas are connected to a Fusion Center (FC) via limited-capacity backhaul links. Similar to cloud radio access networks in communications, the receive antennas quantize the received baseband signal before transmitting it to the FC. The problem of maximizing the detection performance at the FC jointly over the code vector used by the transmitting antenna and over the statistics of the noise introduced by backhaul quantization is investigated. Specifically, adopting the information-theoretic criterion of the Bhattacharyya distance to evaluate the detection performance at the FC and information-theoretic measures of the quantization rate, the problem at hand is addressed via a Block Coordinate Descent (BCD) method coupled with Majorization-Minimization (MM). Numerical results demonstrate the advantages of the proposed joint optimization approach over more conventional solutions that perform separate optimization.

Cost analysis of compressive sensing for MIMO STAP random arrays

Abstract: This work proposes an augmented variation of conventional space-time adaptive processing (STAP), and explores the application of multi-branch matching pursuit (MBMP) to a multiple-input multiple-output (MIMO) beamformer whose steering vector is created over an array having random, inter-element spacing. By applying compressive sensing (CS), a radar system is able to minimize the undesired effects of an undersampled array while providing adequate clutter suppression and reduced burden on array integration. In this paper, we compare the performance and computational complexity of the MBMP applied to the STAP problem and the STAP beamformer. In addition we propose two methods to reduce the computational complexity of MBMP, a modification to the MBMP algorithm which we refer to as truncated MBMP, and a grid refinement technique. We evaluate our approach and extend this aspect to help in understanding the necessary computations required for practical target detection.

Modulation Classification for MIMO-OFDM Signals via Approximate Bayesian Inference

Abstract: The problem of modulation classification for a multiple-antenna (MIMO) system employing orthogonal frequency division multiplexing (OFDM) is investigated under the assumption of unknown frequency-selective fading channels and signal-to-noise ratio (SNR). The classification problem is formulated as a Bayesian inference task, and solutions are proposed based on Gibbs sampling and mean field variational inference. The proposed methods rely on a selection of the prior distributions that adopts a latent Dirichlet model for the modulation type and on the Bayesian network formalism. The Gibbs sampling method converges to the optimal Bayesian solution and, using numerical results, its accuracy is seen to improve for small sample sizes when switching to the mean field variational inference technique after a number of iterations. The speed of convergence is shown to improve via annealing and random restarts. While most of the literature on modulation classification assume that the channels are flat fading, that the number of receive antennas is no less than that of transmit antennas, and that a large number of observed data symbols are available, the proposed methods perform well under more general conditions. Finally, the proposed Bayesian methods are demonstrated to improve over existing non-Bayesian approaches based on independent component analysis and on prior Bayesian methods based on the `superconstellation' method.

Optimization of multistatic cloud radar with multiple-access wireless backhaul

Abstract: A multistatic cloud radar system is investigated, where receive antennas (RAs), or sensors, communicate with a fusion center (FC) over a multiple-access wireless backhaul. Each RA receives a measurement of the signal sent by a transmit antenna (TA) and reflected from target, possibly in the presence of clutter and interference, amplifies it, and forwards it to the FC on a wireless fading channel. The FC receives the signals transmitted by the RAs and determines the presence of a target. The problem of maximizing the Bhattacharyya distance as the detection performance metric under power constraints for the TA and RAs is formulated with respect to the transmitted code vector and the gains applied at the RAs. A short-term adaptive design is first considered that leverages the instant gain of the RAs-to-FC channels, and then a long-term adaptive design is considered that uses only stochastic channel state information (CSI). Algorithmic solutions for both scenarios are proposed based on successive convex approximation, and the performance is evaluated via numerical results.

High Precision TOA-based Direct Localization of Multiple Sources in Multipath

Abstract: Localization of radio frequency sources over multipath channels is a difficult problem arising in applications such as outdoor or indoor gelocation. Common approaches that combine ad-hoc methods for multipath mitigation with indirect localization relying on intermediary parameters such as time-of-arrivals, time difference of arrivals or received signal strengths, provide limited performance. This work models the localization of known waveforms over unknown multipath channels in a sparse framework, and develops a direct approach in which multiple sources are localized jointly, directly from observations obtained at distributed sources. The proposed approach exploits channel properties that enable to distinguish line-of-sight (LOS) from non-LOS signal paths. Theoretical guarantees are established for correct recovery of the sources' locations by atomic norm minimization. A second-order cone-based algorithm is developed to produce the optimal atomic decomposition, and it is shown to produce high accuracy location estimates over complex scenes, in which sources are subject to diverse multipath conditions, including lack of LOS.

Modulation classification for MIMO-OFDM signals via Gibbs sampling

Abstract: The problem of modulation classification for a multiple-antenna (MIMO) system employing orthogonal frequency division multiplexing (OFDM) is investigated under the assumptions of unknown frequency-selective fading channels and signal-to-noise ratio (SNR). The classification problem is formulated as a Bayesian inference task and a solution is proposed based on a selection of the prior distributions that adopts a latent Dirichlet model for the modulation type and on the Bayesian network formalism. The proposed Gibbs sampling method converges to the optimal Bayesian solution and the speed of convergence is shown to improve via annealing and random restarts. While most of the existing modulation classification techniques works under the assumptions that the channels are flat fading and that a large amount of observed data symbols is available, the proposed approach performs well under more general conditions. Finally, the proposed Bayesian method is demonstrated to improve over existing non-Bayesian approaches based on independent component analysis.

Multistatic Cloud Radar Systems: Joint Sensing and Communication Design

Abstract: In a multistatic cloud radar system, receive sensors measure signals sent by a transmit element and reflected from a target and possibly clutter, in the presence of interference and noise. The receive sensors communicate over non-ideal backhaul links with a fusion center, or cloud processor, where the presence or absence of the target is determined. The backhaul architecture can be characterized either by an orthogonal-access channel or by a non-orthogonal multiple-access channel. Two backhaul transmission strategies are considered, namely compress-and-forward (CF), which is well suited for the orthogonal-access backhaul, and amplify-and-forward (AF), which leverages the superposition property of the non-orthogonal multiple-access channel. In this paper, the joint optimization of the sensing and backhaul communication functions of the cloud radar system is studied. Specifically, the transmitted waveform is jointly optimized with backhaul quantization in the case of CF backhaul transmission and with the amplifying gains of the sensors for the AF backhaul strategy. In both cases, the information-theoretic criterion of the Bhattacharyya distance is adopted as a metric for the detection performance. Algorithmic solutions based on successive convex approximation are developed under different assumptions on the available channel state information (CSI). Numerical results demonstrate that the proposed schemes outperform conventional solutions that perform separate optimizations of the waveform and backhaul operation, as well as the standard distributed detection approach.

Theoretical bound on modulation classification for multiple-input multiple-output (MIMO) systems over unknown, flat fading channels

Abstract: Likelihood-based algorithms identify the modulation of the transmitted signal based on the computation of the likelihood function of received signals under different hypotheses (modulation formats). An important class of likelihood-based algorithms for modulation classification problems first treats the unknown channels as deterministic, and replaces the channels by their estimates. In this paper, a novel theoretical bound on the performance of this class of algorithms is proposed for multiple-input multiple-output (MIMO) systems over unknown, flat fading channels. The performance bound is developed from the Cramer-Rao bound (CRB) of blind channel estimation. It provides a useful benchmark against which it is possible to compare the performance of modulation classification algorithms, and is tighter than the theoretical bound derived based on perfect channel knowledge.

Positioning via Direct Localization in C-RAN Systems

Abstract: Cloud Radio Access Network (C-RAN) is a prominent architecture for 5G wireless cellular system that is based on the centralization of baseband processing for multiple distributed radio units (RUs) at a control unit (CU). In this work, it is proposed to leverage the C-RAN architecture to enable the implementation of direct localization of the position of mobile devices from the received signals at distributed RUs. With ideal connections between the CU and the RUs, direct localization is known to outperform traditional indirect localization, whereby the location of a source is estimated from intermediary parameters estimated at the RUs. However, in a C-RAN system with capacity limited fronthaul links, the advantage of direct localization may be offset by the distortion caused by the quantization of the received signal at the RUs. In this paper, the performance of direct localization is studied by accounting for the effect of fronthaul quantization with or without dithering. An approximate Maximum Likelihood (ML) localization is developed. Then, the Cramer-Rao Bound (CRB) on the squared position error (SPE) of direct localization with quantized observations is derived. Finally, the performance of indirect localization and direct localization with or without dithering is compared via numerical results.

Effect of MAC type and speed on neighbor discovery in wireless train networks

Abstract: In this paper, a wireless neighbor discovery (ND) process in a linear topology of a high-speed train backbone network is studied. A key step that enables communication in such a network is that of topology discovery (TD), whereby nodes learn in a distributed fashion the physical topology of the backbone network. ND, where each individual node discovers their right and left one-hop neighbors is the first and key step in TD. While the current standard for train inauguration assumes wired links between adjacent backbone nodes, this paper investigates the more challenging scenario in which the nodes communicate in wireless-fashion based on IEEE 802.11 or slotted-ALOHA. A network simulation using NS-2 software is developed for the 802.11-based (and slotted-ALOHA based) wireless ND. The network simulation is applied to analyze performance metrics, such as ND time and ND success rate as a function of various parameters.

Multistatic Cloud Radar Systems: Joint Waveform and Backhaul Optimization.

Abstract: In a multistatic cloud radar system, receive elements (REs) measure signals sent by a transmit element (TE) and reflected from a target and possibly clutter, in the presence of interference and noise. The REs communicate over non-ideal backhaul links with a fusion center (FC), or cloud processor, where the presence or absence of the target is determined. Two different backhaul architectures are considered, namely orthogonal-access and multiple-access backhaul. For the former case, the REs quantize the received baseband signals prior to forwarding them to the FC in order to satisfy the backhaul capacity constraints; instead, in the latter case, the REs amplify and forward the received signals so as to leverage the superposition properties of the backhaul channel. This paper addresses the joint optimization of the sensing and communication functions of the cloud radar system adopting the information-theoretic criterion of the Bhattacharyya distance as a proxy for the detection performance. Specifically, the transmitted waveform is jointly optimized with the backhaul quantization in the case of orthogonal-access and with the amplifying gains of the REs in the case multiple-access backhaul. Algorithmic solutions based on successive convex approximation are developed for instantaneous or stochastic channel state information (CSI) on the REs-to-FC channels. Numerical results demonstrate that the proposed schemes outperform conventional solutions that perform separate optimizations of the waveform and backhaul operation, as well as the standard distributed detection approach.