abstract: Distributed inference has applications in fields as varied as source localization, evaluation of network quality, and remote monitoring of wildlife habitats. In this dissertation, distributed inference algorithms over multiple-access channels are considered. The performance of these algorithms and the effects of wireless communication channels on the performance are studied. In a first class of problems, distributed inference over fading Gaussian multiple-access channels with amplify-and-forward is considered. Sensors observe a phenomenon and transmit their observations using the amplify-and-forward scheme to a fusion center (FC). Distributed estimation is considered with a single antenna at the FC, where the performance is evaluated using the asymptotic variance of the estimator. The loss in performance due to varying assumptions on the limited amounts of channel information at the sensors is quantified. With multiple antennas at the FC, a distributed detection problem is also considered, where the error exponent is used to evaluate performance. It is shown that for zero-mean channels between the sensors and the FC when there is no channel information at the sensors, arbitrarily large gains in the error exponent can be obtained with sufficient increase in the number of antennas at the FC. In stark contrast, when there is channel information at the sensors, the gain in error exponent due to having multiple antennas at the FC is shown to be no more than a factor of 8/π for Rayleigh fading channels between the sensors and the FC, independent of the number of antennas at the FC, or correlation among noise samples across sensors. In a second class of problems, sensor observations are transmitted to the FC using constant-modulus phase modulation over Gaussian multiple-access-channels. The phase modulation scheme allows for constant transmit power and estimation of moments other than the mean with a single transmission from the sensors. Estimators are developed for the mean, variance and signal-to-noise ratio (SNR) of the sensor observations. The performance of these estimators is studied for different distributions of the observations. It is proved that the estimator of the mean is asymptotically efficient if and only if the distribution of the sensor observations is Gaussian. / Dissertation/Thesis / Ph.D. Electrical Engineering 2010

In order to better identify and understand the differences between Japanese and English, the task of giving walking directions was used. Japanese and American public facilities (10 each) were randomly chosen from which to collect data over the phone in order to examine these differences based on the following five communication styles: 1) politeness, 2) indirectness, 3) self-effacement, 4) back-channel feedback (Aizuchi), 5) and other linguistic and cognitive differences in relation to space and giving directions. The study confirmed what the author reviewed in the literature: Japanese are more polite, English speakers tend to give directions simply and precisely, Japanese prefer pictorial information and most Americans prefer linguistic information, Japanese is a topic-oriented language and also an addressee-oriented language. The information revealed from this study will help Japanese learners develop important skills needed for developing proficiency in the target language and also teach important differences between the two languages.

This dissertation designs linear transceivers for the multiuser downlink in multiple-input multiple-output (MIMO) systems. The designs rely on an uplink/downlink duality for the mean squared error (MSE) of each individual data stream. We first consider the design of transceivers assuming channel state information (CSI) at the transmitter. We consider minimization of the sum-MSE over all users subject to a sum power constraint on each transmission. Using MSE duality, we solve a computationally simpler convex problem in a virtual uplink. The transformation back to the downlink is simplified by our demonstrating the equality of the optimal power allocations in the uplink and downlink. Our second set of designs maximize the sum throughput for all users. We establish a series of relationships linking MSE to the signal-to-interference-plus-noise ratios of individual data streams and the information theoretic channel capacity under linear minimum MSE decoding. We show that minimizing the product of MSE matrix determinants is equivalent to sum-rate maximization, but we demonstrate that this problem does not admit a computationally efficient solution. We simplify the problem by minimizing the product of mean squared errors (PMSE) and propose an iterative algorithm based on alternating optimization with near-optimal performance. The remainder of the thesis considers the more practical case of imperfections in CSI. First, we consider the impact of delay and limited-rate feedback. We propose a system which employs Kalman prediction to mitigate delay; feedback rate is limited by employing adaptive delta modulation. Next, we consider the robust design of the sum-MSE and PMSE minimizing precoders with delay-free but imperfect estimates of the CSI. We extend the MSE duality to the case of imperfect CSI, and consider a new optimization problem which jointly optimizes the energy allocations for training and data stages along with the sum-MSE/PMSE minimizing transceivers. We prove the separability of these two problems when all users have equal estimation error variances, and propose several techniques to address the more challenging case of unequal estimation errors.

This dissertation designs linear transceivers for the multiuser downlink in multiple-input multiple-output (MIMO) systems. The designs rely on an uplink/downlink duality for the mean squared error (MSE) of each individual data stream. We first consider the design of transceivers assuming channel state information (CSI) at the transmitter. We consider minimization of the sum-MSE over all users subject to a sum power constraint on each transmission. Using MSE duality, we solve a computationally simpler convex problem in a virtual uplink. The transformation back to the downlink is simplified by our demonstrating the equality of the optimal power allocations in the uplink and downlink. Our second set of designs maximize the sum throughput for all users. We establish a series of relationships linking MSE to the signal-to-interference-plus-noise ratios of individual data streams and the information theoretic channel capacity under linear minimum MSE decoding. We show that minimizing the product of MSE matrix determinants is equivalent to sum-rate maximization, but we demonstrate that this problem does not admit a computationally efficient solution. We simplify the problem by minimizing the product of mean squared errors (PMSE) and propose an iterative algorithm based on alternating optimization with near-optimal performance. The remainder of the thesis considers the more practical case of imperfections in CSI. First, we consider the impact of delay and limited-rate feedback. We propose a system which employs Kalman prediction to mitigate delay; feedback rate is limited by employing adaptive delta modulation. Next, we consider the robust design of the sum-MSE and PMSE minimizing precoders with delay-free but imperfect estimates of the CSI. We extend the MSE duality to the case of imperfect CSI, and consider a new optimization problem which jointly optimizes the energy allocations for training and data stages along with the sum-MSE/PMSE minimizing transceivers. We prove the separability of these two problems when all users have equal estimation error variances, and propose several techniques to address the more challenging case of unequal estimation errors.

La technologie OFDMA (Orthogonal frequency division multiple access) a été adoptée par les systèmes de télécommunications de 4ème génération (4G) comme technique de transmission et d'accès multiple pour ses performances supérieures en termes d'efficacité spectrale. Dans ce type de systèmes, l'adaptation dynamique du débit en fonction de la qualité du canal CQI (Channel Quality Indicator) constitue une problématique de recherche d'actualité qui attire l'attention de plusieurs acteurs académiques et industriels. Ce problème d'adaptation dynamique est encore plus complexe à gérer dans des environnements multi-utilisateurs hétérogènes et à ressources limitées tels que les systèmes OFDMA comme WiMAX Mobile et Long-term Evolution (LTE). Dans cette thèse, nous nous intéressons au problème d'allocation de ressources de l'information de feedback relative au CQI dans le cadre de systèmes OFDMA multi-porteuses multi-utilisateurs. Dans le but de réduire la charge (overhead) du feedback, nous proposons une méthode de prédiction du CQI basée sur l'exploitation de la corrélation temporelle de ce dernier et d'une solution inter-couches. L'objectif est de trouver des schémas d'allocation de ressources adaptatifs respectant les contraintes de qualité de service (QoS) applicatives.Nous proposons en premier lieu un algorithme de réduction de feedback PBF (Prediction Based Feedack) qui permet à la station de base (BS) à prédire certaines occurrences du CQI en se basant sur l'algorithme des moindres carrés récursif RLS (Recursive least-square). Les résultats de simulation montrent que l'outil de prédiction du CQI réduit sensiblement l'overhead du feedback et améliore par conséquent le débit de la liaison montante. Nous proposons, par la suite, une version opportuniste de PBF pour atténuer les éventuels effets de sur et sous estimations liées à l'algorithme de prédiction. Dans ce mécanisme, nous exploitons les informations inter-couches pour améliorer les performances des mécanismes de feedbacks périodiques dont PBF fait partie. L'approche opportuniste améliore sensiblement les performances du système pour les cas de mobilité élevée comparés aux cas de faible mobilité.Dans un second temps, nous proposons une plateforme (FEREP : feedback resource allocation and prediction) basée sur une approche inter-couches. Implémentée au niveau de la station BS, FEREP intègre les fonctionnalités de prédiction, d'adaptation dynamique du CQI et d'ordonnancement des demandes de feedback. Elle comporte trois modules. Le module FWA (feedback window adaptation) gère dynamiquement la fenêtre de feedbacks de chaque station mobile (MS) en se basant sur les messages ARQ (Automatic Repeat Request) reçus qui reflètent l'état actuel des canaux respectifs. Le module PBFS (priority-based feedback scheduling) effectue ensuite l'ordonnancement des feedbacks en tenant compte de la taille de la fenêtre de feedback, du profil de l'utilisateur sous la contrainte de la limitation des ressources globales du systèmes réservées au feedback. Afin de choisir les paramètres de transmission MCS (modulation and coding schemes), le module PBF (prediction based feedback) est utilisé pour les utilisateurs dont le feedabck n'a pas pu être ordonnancé dans la trame courante. Les résultats de simulation ont montré un gain significatif des performances de FREREP en comparaison à un mécanisme de référence, en particulier, sous de fortes contraintes de limitation des ressources du feedback.Le protocole ARQ génère un accusé de réception uniquement si l'utilisateur est sélectionné par l'ordonnanceur pour envoyer des données sur la liaison descendante. Dans le cas où la fréquence d'ordonnancement des utilisateurs sur le lien descendant est réduite, les messages ARQ s'en trouvent également réduits, dégradant par conséquent les performances de la plateforme FEREP proposée ci-dessus. En effet, dans ce cas la signalisation ARQ devient insuffisante pour adapter efficacement la fenêtre de feedback de chaque utilisateur. Pour pallier à ce problème, nous proposons l'algorithme DCRA (dynamic CQI resource allocation) qui utilise deux modes d'estimation de la fenêtre de feedback. Le premier est un mode hors-ligne basé sur des études empiriques permettant d'estimer la fenêtre moyenne optimale de feedback en utilisant les profils applicatif et de mobilité de l'utilisateur. Notre analyse de performance par simulation montre que la fenêtre de feedback peut être estimée en fonction de la classe de service des utilisateurs et de leurs profils de mobilité pour un environnement cellulaire donné. Le second mode de fonctionnement de DCRA effectue une adaptation dynamique de la fenêtre en temps réel dans le cas où la signalisation ARQ est suffisante. Une étude comparative avec les mécanismes DFS (deterministic feedback scheduling) et OFS (opportunistic feedback scheduling), a montré que DCRA arrive à réaliser un meilleur gain en ressources montantes grâce à la réduction de l'overhead des feedbacks, sans pour autant trop dégrader le débit descendant des utilisateurs. Du point de vue des utilisateurs, DCRA améliore les contraintes de QoS tels que le taux de perte de paquets et réduit la consommation énergétique des terminaux grâce à la réduction de feedback. / Orthogonal frequency division multiple access (OFDMA) technology has been adopted by 4th generation (a.k.a. 4G) telecommunication systems to achieve high system spectral efficiency. A crucial research issue is how to design adaptive channel quality indicator (CQI) feedback mechanisms so that the base station can use adaptive modulation and coding (AMC) techniques to adjust its data rate based on the channel condition. This problem is even more challenging in resource-limited and heterogeneous multiuser environments such as Mobile WiMAX, Long-term Evolution (LTE) networks. In this thesis, we consider CQI feedback resource allocation issue for multiuser multicarrier OFDMA systems. We exploit time-domain correlation for CQI prediction and cross-layer information to reduce feedback overhead for OFDMA systems. Our aim is find resource allocation schemes respecting the users QoS constraints.Our study begins with proposing prediction based feedback (PBF) which allows the base station to predict the CQI feedbacks based on recursive least-square (RLS) algorithm. We showed that it is useful to use channel prediction as a tool to reduce the feedback overhead and improve the uplink throughput. Then, we propose an opportunistic periodic feedback mechanism to mitigate the possible under and over estimation effects of CQI prediction. In this mechanism, we exploited the cross-layer information to enhance the performance of periodic feedback mechanisms. The opportunistic mechanism improves the system performance for high mobility cases compared to low mobility cases.For OFDMA systems with limited feedback resource, we propose an integrated cross-layer framework of feedback resource allocation and prediction (FEREP). The proposed framework, implemented at the BS side, is composed of three modules. The feedback window adaptation (FWA) module dynamically tunes the feedback window size for each mobile station based on the received ARQ (Automatic Repeat Request) messages that reflect the current channel condition. The priority-based feedback scheduling (PBFS) module then performs feedback allocation by taking into account the feedback window size, the user profile and the total system feedback budget. To choose adapted modulation and coding schemes (MCS), the prediction based feedback (PBF) module performs channel prediction by using recursive least square (RLS) algorithm for the user whose channel feedback has not been granted for schedule in current frame. Through extensive simulations, the proposed framework shows significant performance gain especially under stringent feedback budget constraint.ARQ protocol receives users acknowledgement only if the user is scheduled in the downlink. The reduction in users scheduling frequency also reduces the rate of ARQ hints and degrades the performance of above contributions. In this case, it is difficult to exploit the ARQ signal to adapt the feedback window for that user. To address this issue, we propose a cross-layer dynamic CQI resource allocation (DCRA) algorithm for multiuser multicarrier OFDMA systems. DCRA uses two modes for feedback window estimation. The first one is an off-line mode based on empirical studies to derive optimal average feedback window based on user application and mobility profile. Our experimental analysis shows that the feedback window can be averaged according to users service class and their mobility profile for a given cell environment. DCRA performs a realtime dynamic window adaptation if sufficient cross-layer hints are available from ARQ signaling. DCRA increases uplink resource by reducing feedback overhead without degrading downlink throughout significantly compared to deterministic feedback scheduling (DFS) and opportunistic feedback scheduling (OFS). From the users perspective, DCRA improves QoS constraints like packet loss rate and saves users power due to feedback reduction.