Search

A transmission scheme for mean square stabilization of two linear systems over a Gaussian interference relay channel is studied. A delay-free linear sensing and control strategy is proposed and an achievable stability region is derived. It shows that the stability region can be significantly enlarged by deploying a relay node in such a multi-user Gaussian channels. Furthermore we observe that the separation structure between estimation and control is inadequate in high interference regime., QC 20140121

In this paper, we propose two algorithms to construct motion-adaptive transforms that are based on vertex-weighted graphs. The graphs are constructed by motion vector information. The weights of the vertices are given by scale factors that are used to accommodate proper concentration of energy in transforms. The vertex-weighted graph defines a one dimensional linear subspace. Thus, our transform basis is subspace constrained. We propose two algorithms. The first is based on the Gram-Schmidt orthonormalization of the discrete cosine transform (DCT) basis. The second combines the rotation of the DCT basis and the Gram-Schmidt orthonormalization. We assess both algorithms in terms of energy compaction. Moreover, we compare to prior work on graph-based rotation of the DCT basis and on so-called motion-compensated orthogonal transforms (MCOT). In our experiments, both algorithms outperform MCOT in terms of energy compaction. However, their performance is similar to that of graph-based rotation of the DCT basis., QC 20140312

This paper investigates amplify-and-forward (AF) relaying for secrecy in quasi-static Rayleigh fading channels. We consider a four-node network where a helping node intends to enhance secrecy of the transmission between the source and the destination in presence of a passive eavesdropper. In this scenario, the common assumption of full CSI on the eavesdropper's channels is not realistic, and thus, we study the performance of cooperation from an outage perspective. Starting from the secrecy outage probability, we introduce a novel measure, the conditional secrecy outage probability to analyze the performance of AF. In particular, we derive closed-form expressions for AF for these two secrecy measures under a high SNR assumption. Moreover, we use numerical examples to illustrate our results and to characterize the effect of the nodes' geometry. We also show numerically how AF improves the secrecy performance in comparison to direct transmission in terms of outage probability and secure throughput., QC 20131112

In this work, we study the so-called leakage minimization problem, within the context of interference alignment (IA). For that purpose, we propose a novel approach based on controlled perturbations of the leakage function, and show how the latter can be used as a mechanism to control the algorithm's convergence (and thus tradeoff convergence speed for reliability). Although the proposed scheme falls under the broad category of stochastic optimization, we show through simulations that it has a quasi-deterministic convergence that we exploit to improve on the worst case performance of its predecessor, resulting in significantly better sum-rate capacity and average cost function value., QC 20140602, METIS

In this work, we study the so-called leakage minimization problem, within the context of interference alignment (IA). For that purpose, we propose a novel approach based on controlled perturbations of the leakage function, and show how the latter can be used as a mechanism to control the algorithm's convergence (and thus tradeoff convergence speed for reliability). Although the proposed scheme falls under the broad category of stochastic optimization, we show through simulations that it has a quasi-deterministic convergence that we exploit to improve on the worst case performance of its predecessor, resulting in significantly better sum-rate capacity and average cost function value., QC 20140602, METIS

In this work, we study the so-called leakage minimization problem, within the context of interference alignment (IA). For that purpose, we propose a novel approach based on controlled perturbations of the leakage function, and show how the latter can be used as a mechanism to control the algorithm's convergence (and thus tradeoff convergence speed for reliability). Although the proposed scheme falls under the broad category of stochastic optimization, we show through simulations that it has a quasi-deterministic convergence that we exploit to improve on the worst case performance of its predecessor, resulting in significantly better sum-rate capacity and average cost function value., QC 20140602, METIS

In this work, we study the so-called leakage minimization problem, within the context of interference alignment (IA). For that purpose, we propose a novel approach based on controlled perturbations of the leakage function, and show how the latter can be used as a mechanism to control the algorithm's convergence (and thus tradeoff convergence speed for reliability). Although the proposed scheme falls under the broad category of stochastic optimization, we show through simulations that it has a quasi-deterministic convergence that we exploit to improve on the worst case performance of its predecessor, resulting in significantly better sum-rate capacity and average cost function value., QC 20140602, METIS

In this work, we study the so-called leakage minimization problem, within the context of interference alignment (IA). For that purpose, we propose a novel approach based on controlled perturbations of the leakage function, and show how the latter can be used as a mechanism to control the algorithm's convergence (and thus tradeoff convergence speed for reliability). Although the proposed scheme falls under the broad category of stochastic optimization, we show through simulations that it has a quasi-deterministic convergence that we exploit to improve on the worst case performance of its predecessor, resulting in significantly better sum-rate capacity and average cost function value., QC 20140602, METIS

Operation-intensive exhaustive checking is the unique optimal method for performing joint maximum-likelihood (JML) data and channel estimation. Thus, several suboptimal methods incorporating conventional maximum-likelihood sequence estimation (MLSE) or its extensions have been proposed in order to reduce complexity. However, most of them suffer from either long convergence delay or unsatisfactory performance. Therefore, based on a two-phase MLSE algorithm designed for partially coherent receivers, we propose a novel low-complexity and low-latency batch algorithm for JML process. The proposed algorithm first searches for several of the best candidate training sequences using priority-first search, and then applies the training sequences to the two-phase MLSEs to obtain the maximum-likelihood data sequence. Simulation results show that when the candidate training sequence length is adequate, the proposed algorithm can achieve the performance of optimal exhaustive checking., QC 20140121

Base station cooperation in cellular networks has been recently recognized as a key technology for mitigating interference, providing thus significant improvements in the system performance. In this paper, we consider a simple scenario consisting of two one-dimensional cells, where the base stations have fixed locations, while the user terminals are randomly distributed on a line. Exploiting the replica method from statistical physics, we derive the ergodic sum-rate under arbitrary signaling for both cooperative and non-cooperative scenarios, when the system size grows large. The obtained results are analytically tractable and can be used to optimize the system parameters in a simple manner. The numerical examples show that the analysis provides good approximations for finite-sized systems., QC 20140131

High quality view synthesis is a prerequisite for future free-viewpointtelevision. It will enable viewers to move freely in a dynamicreal world scene. Depth image based rendering algorithms willplay a pivotal role when synthesizing an arbitrary number of novelviews by using a subset of captured views and corresponding depthmaps only. Usually, each depth map is estimated individually bystereo-matching algorithms and, hence, shows lack of inter-viewconsistency. This inconsistency affects the quality of view synthesis negatively. This paper enhances the inter-view consistency ofmultiview depth imagery. First, our approach classifies the colorinformation in the multiview color imagery by modeling color witha mixture of Dirichlet distributions where the model parameters areestimated in a Bayesian framework with variational inference. Second, using the resulting color clusters, we classify the correspondingdepth values in the multiview depth imagery. Each clustered depthimage is subject to further sub-clustering. Finally, the resultingmean of each sub-cluster is used to enhance the depth imagery atmultiple viewpoints. Experiments show that our approach improvesthe average quality of virtual views by up to 0.8 dB when comparedto views synthesized by using conventionally estimated depth maps., QC 20140224

This note proves that, for any finite set of correlated discrete i.i.d. sources, there always exists a sequence of linear encoders over some finite non-field rings which achieves the data compression limit, the Slepian-Wolf region. Based on this, we address a variation of the data compression problem which considers recovering some discrete function of the data. It is demonstrated that linear encoder over non-field ring strictly outperforms its field counterpart for encoding some function in terms of achieving strictly larger achievable region with strictly smaller alphabet size., QC 20140311

We consider a secure source coding problem with side informations at the decoder and the eavesdropper. The encoder has a source that it wishes to describe with limited distortion through a rate-limited link to a legitimate decoder. The message sent is also observed by the eavesdropper. The encoder aims to minimize both the distortion incurred by the legitimate decoder; and the information leakage rate at the eavesdropper. When the encoder has access to the side information (S.I.) at the decoder, we characterize the rate-distortion-information leakage rate (R.D.I.) region under a Markov chain assumption and when S.I. at the encoder does not improve the rate-distortion region as compared to the case when S.I. is absent. We then extend our setting to consider the case where the encoder and decoder obtain coded S.I. through a rate-limited helper, and characterize the R.D.I. region for several special cases under logarithmic loss distortion (log-loss). Finally, we consider the case of list or entropy constraints at the decoder and show that the R.D.I. region coincides with R.D.I. region under log-loss., QC 20140312

The problem of feedback stabilization of LTI plants over a Gaussian interference channel is considered. Two plants with arbitrary distributed initial states are monitored by two separate sensors which communicate their measurements to two separate controllers over a Gaussian interference channel under average transmit power constraints. The necessary conditions for mean square-stabilization over a memoryless symmetric Gaussian interference channel are derived. These conditions are shown to be tight for some system parameters. Further it is shown that linear memoryless sensing and control schemes are optimal for stabilization in some special cases., QC 20140128

We study the achievable sum degrees of freedom(DoF) in a class of wireless single-antenna multi-hop half-duplexrelay networks. The networks contain Ms information sources, Md information destinations, and arbitrary layers of relays, each with 2K (K ≥ max{Ms,Md}) half-duplex relays, in between. A cluster successive relaying transmission scheme is applied to conduct the communication: We divide the relays in each layer into two equal-size clusters and activate them successively to efficiently use the available channel. It is shown that in a time-varying fading environment this scheme asymptotically achieves the sum DoF min{MsK/(Ms+K) , MdK/(Md+K−1)}, which is irrelevant to the number of hops the source messages have to pass through. This result also implies that, when the number of relays in each layeris infinitely large, the available DoF (i.e. the optimally achievable sum DoF) of the considered networks is min{Ms,Md}. Neither distributed signal processing nor multiple layers of half-duplex relay operation negatively affects the system DoF performance., QC 20130903

We consider a downlink two-hop MISO broadcast network with a 2-antenna source communicating to 2 single-antenna destinations, via 2 single-antenna relays. We investigate spectrally efficient transmission and the associated achievable sum degrees of freedom (DoF) with mixed channel state information at the transmitter (CSIT), which consists of perfect delayed CSIT and imperfect instantaneous CSIT. When the variance of the estimation error of the instantaneous CSIT lies on level of O(P-α) for the transmission power P and some α [0, 1], we show that the sum DoF 4-2α over 3-2α [4 over 3, 2] can be achieved by a novel interference alignment (IA) scheme. The result shows that rather than exploiting only delayed or imperfect instantaneous CSIT, the transmission design taking advantages of both can achieve higher sum DoF., QC 20140912

We study the low SNR scaling of the non-coherent secret-key agreement capacity over a reciprocal, block-fading channel. For the restricted class of strategies, where one of the nodes is constrained to transmit pilot-only symbols, we show that the secret-key capacity scales as SNR ·log T if T ≤ 1/SNR, where T denotes the coherence period, and as SNR·log(1/SNR) otherwise. Our upper bound is inspired by the genie-aided argument of Borade and Zheng (IT-Trans 2010). Our lower bound is based on bursty communication, channel training, and secret message transmission., QC 20140121

Collecting sensory data at access point (AP) from large number of sensor nodes with low latency is a critical issue. In Wi-Fi, prior to uplink data delivery, AP typically needs to poll large number of sensor nodes sequentially and allocate channel resources to individual node resulting in large latency. An efficient method to reduce the latency and power consumption in wireless sensor networks is to parallelize the polling operation so that multiple nodes can concurrently respond to the poll request of an AP by sending orthogonal sequences with uplink power control. In this paper, we present a conceptually simple uplink power control scheme for the parallel polling operation between AP and low power sensor nodes. We formulate the uplink power control problem as a sequence design problem and show that uplink channel state information (CSI) required to achieve a given target receive SNR can be significantly reduced by carefully designing sequences. We further develop a low complexity instantaneous (fast) power control scheme in order to reduce the number of computations required by the low power sensor node. We also analyze and compare the detection performance of the instantaneous (fast) and average (slow) power control schemes in terms of diversity gain., QC 20140317

Reconstruction fidelity of sparse signals contaminated by sparse noise is considered. Statistical mechanics inspired tools are used to show that the l(1)-norm based convex optimization algorithm exhibits a phase transition between the possibility of perfect and imperfect reconstruction. Conditions characterizing this threshold are derived and the mean square error of the estimate is obtained for the case when perfect reconstruction is not possible. Detailed calculations are provided to expose the mathematical tools to a wide audience., QC 20140625

We develop a greedy pursuit algorithm for solving the distributed compressed sensing problem in a connected network. This algorithm is based on subspace pursuit and uses the mixed support-setsignal model. Through experimental evaluation, we show that the distributed algorithm performs significantly better than the standalone (disconnected) solution and close to a centralized (fully connected to a central point) solution., QC 20121114

This paper proposes a new coding scheme for the discrete memoryless interference channel with a dedicated relay. The scheme is built upon rate-splitting encoding, layered noisy network coding, and joint decoding. The result is extended to two Gaussian channels. For the Gaussian channel whose relay is connected to the destinations via orthogonal links we indirectly show that the proposed scheme achieves a bounded gap to the capacity region under certain channel conditions. For the Gaussian channel wherein the relay receives and transmits in the same spectral resource with the transmitters the numerical results show that the proposed scheme achieves higher sum rate than other compress-forward-based schemes. This work, together with our previous work [1], shows that noisy network coding can be extended by the well-known rate-splitting technique of the interference channel to achieve a bounded gap to the capacity region of some multi-unicast networks., QC 20121212

The sparse representation problem of recovering an N dimensional sparse vector x from M < N linear observations y = Dx given dictionary D is considered. The standard approach is to let the elements of the dictionary be independent and identically distributed (IID) zero-mean Gaussian and minimize the l1-norm of x under the constraint y = Dx. In this paper, the performance of l1-reconstruction is analyzed, when the dictionary is bi-orthogonal D = [O1 O2], where O1, O 2 are independent and drawn uniformly according to the Haar measure on the group of orthogonal M × M matrices. By an application of the replica method, we obtain the critical conditions under which perfect l 1-recovery is possible with bi-orthogonal dictionaries., QC 20130219

We propose a LDPC Convolutional Code ensemble together with an expanding-window message-passing decoder that asymptotically have anytime properties when used for streaming transmission on the binary erasure channel. We show analytically that the decoding erasure probability of these codes decays exponentially over decoding delay and determine the corresponding anytime exponents., QC 20130118

This paper explores the trade-off between cooperation and secrecy in cognitive radio networks. We consider a scenario consisting of a primary and a secondary system. In the simplest case, each system is represented by a pair of transmitter and receiver. We assume a secrecy constraint on the transmission in the sense that the message of the primary transmitter has to be concealed from the secondary receiver. Both situations where the secondary transmitter is aware and unaware of the primary message are investigated and compared. In the first case, the secondary transmitter helps by allocating power for jamming, which increases the secrecy of the first message. In the latter case, it can also act as a relay for the primary message, thus improving the reliability of the primary transmission. Furthermore, we extend our results to the scenario where the secondary system comprises multiple receivers. For each case we present achievable rate regions. We then provide numerical illustrations for these rate regions. Our main result is that, in spite of the secrecy constraint, cooperation is beneficial in terms of the achievable rates. In particular, the secondary system can achieve a significant rate without decreasing the primary rate below the benchmark rate achievable without the help of the secondary transmitter. Finally, we investigate the influence of the distances between users on the system's performance., QC 20121219

We analyze the traffic carrying capacity of a single-input single-output system for a given delay guarantee at the flow level. A continuous time ON/OFF data source is considered as the traffic model while the wireless channel is modeled using the Gilbert-Elliot model. We study the impact of the delay bound violation probability, traffic burstiness, channel memory and the delay requirements at the network layer on the throughput under different signal strength., QC 20130111

We investigate the design and performance of spatially-coupled rateless codes. A modified encoding process is introduced for spatially-coupled Luby Transform (SCLT) codes which leads to an almost regular variable-node degree distribution at the encoding graph. The proposed SCLT codes outperform its counterparts significantly over binary erasure channels, particularly in the erasure floor region. To further improve the erasure floor performance, the approach of spatial coupling is then extended to Raptor codes by concatenating a high-rate pre-coder to the SCLT codes. It is shown that the spatial coupling improves the convergence threshold of Raptor codes. Different ensembles of spatially-coupled Raptor codes are constructed depending on whether pre-coders and/or LT codes are spatially-coupled. The performance of different ensembles of spatially-coupled Raptor codes is then evaluated and compared based on density evolution, leading to an improved spatially-coupled Raptor code in terms of convergence threshold and lower complexity., QC 20130115

We consider the problem of detecting a sparse random signal from the compressive measurements without reconstructing the signal. Using a subspace model for the sparse signal where the signal parameters are drawn according to Gaussian law, we obtain the detector based on Neyman-Pearson criterion and analytically determine its operating characteristics when the signal covariance is known. These results are extended to situations where the covariance cannot be estimated. The results can be used to determine the number of measurements needed for a particular detector performance and also illustrate the presence of an optimal support for a given number of measurements., QC 20121119

Dynamic decode-and-forward (DDF) is an improved decode-and-forward (DF) protocol under which the relay decides based on its received channel-state information (CSI) when to switch from listening mode to transmission mode without knowing the CSI of other links. In this paper we propose to apply rate-compatible LDPC convolutional (RC-LDPCC) codes to the DDF relay channel. The RC-LDPCC codes are constructed by successive graph extensions, and they have been proved analytically to be capacity achieving over the binary erasure channel. In this paper we show that the RC-LDPCC codes fit well with the DDF relaying, and the regularity of degree distributions simplifies the code optimization. Numerical results in terms of bit erasure rate and achievable rate are provided to evaluate the performance of the system. The results show that the RC-LDPCC codes are able to provide high achievable rates for the DDF relay channel., QC 20120831

This paper shows that feedback does not increase the capacity of the two-receiver discrete memoryless bidirectional broadcast channel., QC 20130118

For compressed sensing, in the framework of greedy search reconstruction algorithms, we introduce the notion of initial support-set. The initial support-set is an estimate given to a reconstruction algorithm to improve the performance of the reconstruction. Furthermore, we classify existing greedy search algorithms as being serial or parallel. Based on this classification and the goal of robustness to errors in the initial support-sets we develop a new greedy search algorithm called FROGS. We end the paper with careful numerical experiments concluding that FROGS perform well compared to existing algorithms (both in terms of performance and execution time) and that it is robust against errors in the initial support-set., QC 20121108

For compressed sensing, in the framework of greedy search reconstruction algorithms, we introduce the notion of initial support-set. The initial support-set is an estimate given to a reconstruction algorithm to improve the performance of the reconstruction. Furthermore, we classify existing greedy search algorithms as being serial or parallel. Based on this classification and the goal of robustness to errors in the initial support-sets we develop a new greedy search algorithm called FROGS. We end the paper with careful numerical experiments concluding that FROGS perform well compared to existing algorithms (both in terms of performance and execution time) and that it is robust against errors in the initial support-set., QC 20121108

For compressed sensing, in the framework of greedy search reconstruction algorithms, we introduce the notion of initial support-set. The initial support-set is an estimate given to a reconstruction algorithm to improve the performance of the reconstruction. Furthermore, we classify existing greedy search algorithms as being serial or parallel. Based on this classification and the goal of robustness to errors in the initial support-sets we develop a new greedy search algorithm called FROGS. We end the paper with careful numerical experiments concluding that FROGS perform well compared to existing algorithms (both in terms of performance and execution time) and that it is robust against errors in the initial support-set., QC 20121108

For compressed sensing, in the framework of greedy search reconstruction algorithms, we introduce the notion of initial support-set. The initial support-set is an estimate given to a reconstruction algorithm to improve the performance of the reconstruction. Furthermore, we classify existing greedy search algorithms as being serial or parallel. Based on this classification and the goal of robustness to errors in the initial support-sets we develop a new greedy search algorithm called FROGS. We end the paper with careful numerical experiments concluding that FROGS perform well compared to existing algorithms (both in terms of performance and execution time) and that it is robust against errors in the initial support-set., QC 20121108

For compressed sensing, in the framework of greedy search reconstruction algorithms, we introduce the notion of initial support-set. The initial support-set is an estimate given to a reconstruction algorithm to improve the performance of the reconstruction. Furthermore, we classify existing greedy search algorithms as being serial or parallel. Based on this classification and the goal of robustness to errors in the initial support-sets we develop a new greedy search algorithm called FROGS. We end the paper with careful numerical experiments concluding that FROGS perform well compared to existing algorithms (both in terms of performance and execution time) and that it is robust against errors in the initial support-set., QC 20121108

We propose Gaussian process dynamical models (GPDMs) as a new, nonparametric paradigm in acoustic models of speech. These use multidimensional, continuous state-spaces to overcome familiar issues with discrete-state, HMM-based speech models. The added dimensions allow the state to represent and describe more than just temporal structure as systematic differences in mean, rather than as mere correlations in a residual (which dynamic features or AR-HMMs do). Being based on Gaussian processes, the models avoid restrictive parametric or linearity assumptions on signal structure. We outline GPDM theory, and describe model setup and initialization schemes relevant to speech applications. Experiments demonstrate subjectively better quality of synthesized speech than from comparable HMMs. In addition, there is evidence for unsupervised discovery of salient speech structure., QC 20120308, LISTA

We propose Gaussian process dynamical models (GPDMs) as a new, nonparametric paradigm in acoustic models of speech. These use multidimensional, continuous state-spaces to overcome familiar issues with discrete-state, HMM-based speech models. The added dimensions allow the state to represent and describe more than just temporal structure as systematic differences in mean, rather than as mere correlations in a residual (which dynamic features or AR-HMMs do). Being based on Gaussian processes, the models avoid restrictive parametric or linearity assumptions on signal structure. We outline GPDM theory, and describe model setup and initialization schemes relevant to speech applications. Experiments demonstrate subjectively better quality of synthesized speech than from comparable HMMs. In addition, there is evidence for unsupervised discovery of salient speech structure., QC 20120308, LISTA

We propose Gaussian process dynamical models (GPDMs) as a new, nonparametric paradigm in acoustic models of speech. These use multidimensional, continuous state-spaces to overcome familiar issues with discrete-state, HMM-based speech models. The added dimensions allow the state to represent and describe more than just temporal structure as systematic differences in mean, rather than as mere correlations in a residual (which dynamic features or AR-HMMs do). Being based on Gaussian processes, the models avoid restrictive parametric or linearity assumptions on signal structure. We outline GPDM theory, and describe model setup and initialization schemes relevant to speech applications. Experiments demonstrate subjectively better quality of synthesized speech than from comparable HMMs. In addition, there is evidence for unsupervised discovery of salient speech structure., QC 20120308, LISTA

We propose Gaussian process dynamical models (GPDMs) as a new, nonparametric paradigm in acoustic models of speech. These use multidimensional, continuous state-spaces to overcome familiar issues with discrete-state, HMM-based speech models. The added dimensions allow the state to represent and describe more than just temporal structure as systematic differences in mean, rather than as mere correlations in a residual (which dynamic features or AR-HMMs do). Being based on Gaussian processes, the models avoid restrictive parametric or linearity assumptions on signal structure. We outline GPDM theory, and describe model setup and initialization schemes relevant to speech applications. Experiments demonstrate subjectively better quality of synthesized speech than from comparable HMMs. In addition, there is evidence for unsupervised discovery of salient speech structure., QC 20120308, LISTA

We propose Gaussian process dynamical models (GPDMs) as a new, nonparametric paradigm in acoustic models of speech. These use multidimensional, continuous state-spaces to overcome familiar issues with discrete-state, HMM-based speech models. The added dimensions allow the state to represent and describe more than just temporal structure as systematic differences in mean, rather than as mere correlations in a residual (which dynamic features or AR-HMMs do). Being based on Gaussian processes, the models avoid restrictive parametric or linearity assumptions on signal structure. We outline GPDM theory, and describe model setup and initialization schemes relevant to speech applications. Experiments demonstrate subjectively better quality of synthesized speech than from comparable HMMs. In addition, there is evidence for unsupervised discovery of salient speech structure., QC 20120308, LISTA

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

This work addresses the problem of cooperative spectrum sensing in cognitive radio networks, focusing on the impact of mobility on performance of cooperative sensing. First, a review of the most recent results on cooperative spectrum sensing is provided, resulting in the identification of measurement correlation and frame error rate in the reporting channel as the main parameters influencing the performance of cooperative sensing schemes. Next, the paper discusses the extension of the analysis to the case of mobile sensors, and determines the set of assumptions made in existing literature when taking into account mobility in sensing. The paper moves then to remove some of such assumptions, by presenting simulation results obtained in presence of realistic models for propagation in the considered area, as well as of a realistic mobility model. A comparison between theoretical derivation and simulation results shows that correlation among measurements taken by different sensors and the selected mobility model may significantly affect the sensing performance., QC 20130128

We investigate the design of optimized input distributions for the two-hop Gaussian relay channel under an average power constraint. In our study, the relay is confined to implement an estimate-and-forward strategy. We demonstrate that the rate achievable by an optimized discrete input distribution can significantly outperform that achievable by a Gaussian codebook. The lower bound obtained by the optimized EF scheme is also used to establish an improved lower bound on the rates achievable by compress-and-forward., QC 20130122

This paper proposes two new coding schemes for the discretememoryless two-way relay channel. The main target is to show thebenefits of compress-forward without Wyner-Ziv binning and oflayered relaying in networks wherein a relay is to help multipledestinations, that may have unequal channel quality and/or haveaccess to different side information. Numerical results for aGaussian channel show that the new coding schemes outperformvariants of compress-forward relaying and offer a good trade-offbetween achievable rates and complexity and decoding delay. The ideacan also be applied to other relay networks., QC 20121115

This paper studies linear coding (LC) techniques in the setting of computing functions of correlated memoryless sources. Instead of linear mappings over finite fields, we consider using linear mappings over finite rings as encoders. It is shown that generally the region c×R, where c ≥ 1 is a constant and R is the Slepian-Wolf (SW) region, is achievable with LC over ring (LCoR) when the function to compute is the identity function. c = 1 if the ring used is a field. Hence, LCoR could be suboptimal in terms of achieving the best coding rates (the SW region) for computing the identity function. In spite of that, the ring version shows several advantages. It is demonstrated that there exists a function that is neither linear nor can be linearized over any finite field. Thus, LC over field (LCoF) does not apply directly for computing such a function unless the polynomial approach [1], [2] is used. On the contrary, such a function is linear over some ring. Using LCoR, an achievable region containing the SW region can be obtained for computing this function. In addition, the alphabet sizes of the encoders are strictly smaller than using LCoF. More interestingly, LCoF is not useful if some special requirement is imposed., QC 20130104

The remote stabilization of a first order linear plant over a wireless channel is studied. The plant is assumed to have an arbitrary distributed initial state and the wireless channel between the plant's sensor and the controller is modeled as a white Gaussian channel subject to an external interference signal. In order to combat the interference a dedicated sensor (relay) node is deployed adjacent to the interferer, which relays the interference information to both the plant's sensor and the controller. The sensor and the controller utilize this information to mitigate interference. We use delay-free linear sensing and control scheme in order to derive sufficient conditions for mean square stability. The achievable stability region significantly enlarges with the relay assisted interference cancelation scheme. Moreover the effect of interference can be completely eliminated if the encoder knows all the future values of the interference., QC 20130218

We consider a problem of constrained multi-stage coding for channels with a rewrite option. It is a natural extension of Weissman's channels with action-dependent states to the multistage coding case where an encoder in each stage observes its own message as well as all previous-stage messages, inputs, and outputs. In addition to decoding all messages at the final stage, the new reconstruction constraint introduced in Sumszyk and Steinberg's information embedding with reversible stegotext is imposed on the problem such that the decoder is required to be able to reconstruct all channel input sequences reliably. The complete characterization of the channel capacity region is given for the two-stage case, while the inner and outer bounds to the capacity regions for the cases of three or more stages are provided. For the two-stage case, a discussion regarding the rate constraint of the message in the second stage is also given in which we can draw a connection to the two-stage coding condition which appears in our previous study on channel with action-dependent state and reversible input., QC 20121119