1,721,510 research outputs found
Spectrum sensing and interference mitigation in cognitive radio networks
No full textOne concept to increase spectral efficiency is dynamic spectrum access (DSA). The secondary users coexist with the primary users in the same radio-frequency spectrum. Such shared usage demands for an interference avoidance or interference mitigation at the secondary users. This is especially challenging due to the limited cooperation between the primary users (PUs) and the secondary users (SUs). Motivated by this fact, the focus of this thesis is a comprehensive study on interference management strategies and the characterization of the achievable performance of secondary systems. Spectrum sensing aims at detecting the presence or absence of the PUs. The main challenge encountered is the high requirement on sensitivity, reliability, and agility, especially in case of incomplete knowledge of the transmission channels. Therefore, the SUs need to efficiently utilize the limited a priori knowledge related to the primary transmission to improve the sensing performance. In this thesis, the generalized likelihood ratio test framework is applied to cooperative sensing problems with an unknown structure of the primary signal space and unknown noise variances at the SUs. The efficiency of the resulting spectrum sensing algorithms is demonstrated as well as the effectiveness in countering the “hidden primary user” problem. Based on limited knowledge related to the primary transmission, the SUs’ transceiver strategies are optimized in order to achieve the tradeoff between improved secondary network throughput and, most critically, constrain the performance loss of the primary transmission. For a single-antenna spectrum sharing system, the power allocation strategies are investigated for the SUs subject to different quality of service constraints on the primary link. Not only optimal and low-complexity near-optimal power allocation strategies are developed, but also the achievable performance of the system is approximately evaluated in closed form. Additionally, for multi-antenna spectrum sharing networks, efficient transceiver optimization strategies are developed under the consideration of imperfect channel state information. The robustness, optimality, and convergence behavior of the different proposed algorithms are quantitatively verified and compared. The essential “cognitive” property of DSA in cognitive radio networks consists of two aspects: the acquirement of the useful information from the environment and the utilization of such information to improve the spectrum efficiency. This is demonstrated with the study of a hybrid paradigm, in which the SU exploits the spectrum sensing and location information to adapt the transmit power level. Compared to the standard paradigms, e.g., opportunistic transmission and spectrum sharing without sensing, the proposed strategies in the hybrid paradigm achieve better performance
A methodology for the development of flexible and efficient wireless communication systems
No full text
Non-stationary dual-polarized radio channels : characterization, modeling, and performance
No full textThe performance of wireless multi-antenna communication systems is heavily influenced by the multiple-input multiple-output (MIMO) channel over which transmission takes place. In order to understand the behavior of transmission schemes and improve their performance, accurate channel models are essential. The statistical modeling of wireless channels has been thoroughly studied in the literature, and there exist accurate models, e.g., the WINNER model. However, current models are very limited regarding the temporal behavior of the channel and the consideration of differently polarized antennas, even though the exploitation of the polarization domain can allow for an increased spectral efficiency as well as reduced antenna array sizes. The aim of this thesis is the investigation and the understanding of the impact of realistic channels for radio communication on the performance of multi-polarized MIMO systems. An analysis is made possible by channel measurements at 2.53 GHz. The wireless channel is inherently non-stationary, i.e., the statistical description of the channel changes over time. Changing channel statistics result in time variations of the performance of wireless communication systems. In order to understand the time evolution of the performance of wireless systems, an investigation of the channel non-stationarity is indispensable. We propose a methodology which yields a case-specific definition of regions inside which the channel can be treated as a stationary random process. The resulting local quasi-stationarity (LQS) regions allow for a maximal performance degradation of selected algorithms due to mismatched, e.g., outdated, channel statistics. Based on this methodology and existing methods, an extensive measurement-based analysis of the non-stationarity of single-polarized (SP) and dual-polarized (DP) MIMO channels is performed. Exemplarily, a channel estimation and a beamforming technique are considered. The non-stationarity analysis reveals that the LQS regions are highly dependent on the considered algorithm. The use of the polarization domain of the channel has the potential to provide performance gains, e.g., in terms of spectral efficiency. In order to enable such gains, understanding the impact of the statistical channel parameters on the performance is essential. To this end, a simple but general spatial channel model is developed. This model is analytically tractable and takes into account the presence of dominant components, e.g., due to a line-of-sight connection. Based on this model, an approximation of the (ergodic) achievable rate of SP and DP MIMO systems is derived, which is a function of correlation matrices of the channel. This allows for a direct evaluation of the impact of typical channel properties on the performance without resorting to extensive numerical evaluations. Furthermore, this work investigates when an SP or a DP MIMO system should be used to maximize the achievable rate. A signal-to-noise ratio (SNR) threshold above which DP MIMO systems outperform SP MIMO systems is obtained. This SNR threshold is characterized analytically, in terms of statistical channel parameters, as well as based on measurements. The analysis shows that this threshold lies at realistic SNR values, while it is usually lower when the channel exhibits stronger dominant components
On the achievable rate of stationary fading channels
No full textIn typical mobile communication systems transmission takes place over a time-varying fading channel. The stochastic channel fading process can assumed to be bandlimited and its realization is usually unknown to the receiver. To allow for a coherent signal detection, the channel fading process is often estimated based on pilot symbols which are periodically inserted into the transmit symbols sequence. The achievable data rate with this approach depends on the dynamics of the channel fading process. For this conventional approach, i.e., performing channel estimation solely based on pilot symbols and using it for coherent detection (synchronized detection) in a second step, bounds on the achievable data rate are known. However, in recent years receiver structures got into the focus of research, where the channel estimation is iteratively enhanced based on the reliability information on data symbols (code-aided channel estimation). For this kind of systems, the bounds on the achievable data rate with synchronized detection based on a solely pilot based channel estimation are no longer valid. The study of the possible performance gain when using such receivers with synchronized detection and a code-aided channel estimation in comparison to synchronized detection in combination with a solely pilot based channel estimation poses also the question on the capacity of stationary fading channels. Although such channels are typical for many practical mobile communication systems, already for the simple case of a Rayleigh flat-fading channel the capacity and the capacity-achieving input distribution are unknown. There exist bounds on the capacity, however, most of them are tight only in a limited SNR regime or rely on a peak power constraint. Thinking of this, in the present thesis various aspects regarding the capacity/achievable data rate of stationary Rayleigh fading channels are treated. First, bounds on the achievable data rate with i.i.d. zero-mean proper Gaussian input symbols, which are capacity-achieving in the coherent case, i.e., in case of perfect channel knowledge at the receiver, are derived. These bounds are tight in the sense that the difference between the upper and the lower bound is bounded for all SNRs. The lower bound converges to the coherent capacity for asymptotically small channel dynamics. Furthermore, these bounds are extended to the case of multiple-input multiple-output (MIMO) channels and to the case of frequency selective channels. The comparison of these bounds on the achievable rate with i.i.d. zero-mean proper Gaussian input symbols to the achievable rate while using receivers with synchronized detection based on a solely pilot based channel estimation already gives an indication on the performance of such conventional receiver structures. However, for systems with receivers based on iterative code-aided channel estimation periodic pilot symbols are still used. Therefore, in a further part of the present work the achievable rate with receivers based on synchronized detection and a code-aided channel estimation is studied. For a specific type of such a receiver an approximate upper bound on the achievable rate is derived. The comparison of this approximate upper bound and the achievable data rate with receivers using synchronized detection based on a solely pilot based channel estimation gives an approximate upper bound on the possible gain by using this kind of code-aided channel estimation in comparison to the conventional receiver using a solely pilot based channel estimation. In this context, it is also shown which part of the mutual information of the transmitter and the receiver is discarded when using the conventional receiver with synchronized detection based on a solely pilot based channel estimation. Concerning the typically applied periodic pilot symbols the question arises if these periodic pilot symbols are optimal from an information theoretic perspective. To address this question, the mutual information between transmitter and receiver is studied for a given discrete signaling set. The optimum input distribution, i.e., the one that maximizes the mutual information when restricting to the given signaling set, is given implicitly based on the Kullback-Leibler distance. Thereon it is shown that periodic pilot symbols are not capacity-achieving in general. However, for practical systems they allow for receivers with small computational complexity
Non-stationary dual-polarized radio channels : characterization, modeling, and performance
No full textThe performance of wireless multi-antenna communication systems is heavily influenced by the multiple-input multiple-output (MIMO) channel over which transmission takes place. In order to understand the behavior of transmission schemes and improve their performance, accurate channel models are essential. The statistical modeling of wireless channels has been thoroughly studied in the literature, and there exist accurate models, e.g., the WINNER model. However, current models are very limited regarding the temporal behavior of the channel and the consideration of differently polarized antennas, even though the exploitation of the polarization domain can allow for an increased spectral efficiency as well as reduced antenna array sizes. The aim of this thesis is the investigation and the understanding of the impact of realistic channels for radio communication on the performance of multi-polarized MIMO systems. An analysis is made possible by channel measurements at 2.53 GHz. The wireless channel is inherently non-stationary, i.e., the statistical description of the channel changes over time. Changing channel statistics result in time variations of the performance of wireless communication systems. In order to understand the time evolution of the performance of wireless systems, an investigation of the channel non-stationarity is indispensable. We propose a methodology which yields a case-specific definition of regions inside which the channel can be treated as a stationary random process. The resulting local quasi-stationarity (LQS) regions allow for a maximal performance degradation of selected algorithms due to mismatched, e.g., outdated, channel statistics. Based on this methodology and existing methods, an extensive measurement-based analysis of the non-stationarity of single-polarized (SP) and dual-polarized (DP) MIMO channels is performed. Exemplarily, a channel estimation and a beamforming technique are considered. The non-stationarity analysis reveals that the LQS regions are highly dependent on the considered algorithm. The use of the polarization domain of the channel has the potential to provide performance gains, e.g., in terms of spectral efficiency. In order to enable such gains, understanding the impact of the statistical channel parameters on the performance is essential. To this end, a simple but general spatial channel model is developed. This model is analytically tractable and takes into account the presence of dominant components, e.g., due to a line-of-sight connection. Based on this model, an approximation of the (ergodic) achievable rate of SP and DP MIMO systems is derived, which is a function of correlation matrices of the channel. This allows for a direct evaluation of the impact of typical channel properties on the performance without resorting to extensive numerical evaluations. Furthermore, this work investigates when an SP or a DP MIMO system should be used to maximize the achievable rate. A signal-to-noise ratio (SNR) threshold above which DP MIMO systems outperform SP MIMO systems is obtained. This SNR threshold is characterized analytically, in terms of statistical channel parameters, as well as based on measurements. The analysis shows that this threshold lies at realistic SNR values, while it is usually lower when the channel exhibits stronger dominant components
Interference mitigation in multicell networks
No full textThis thesis investigates the theoretical and algorithmic framework for intercell interference mitigation in multicell networks based on transmitter optimization. A problem in multiuser multicell communication systems is the unfair distribution of the achievable data rate. By the exchange of channel state information (CSI) and cooperation among base station antenna arrays max--min fairness can be achieved in the network. Especially in multicell networks, the CSI is out-dated very rapidly. Consequently, fast algorithms are required for the optimization. The theory of the unicast max--min beamforming problem (MBP) in a single base station scenario with a single sum power constraint is well understood. Low complexity algorithms based on uplink--downlink duality exist. This thesis extends the duality framework of the single base station scenario to the multicell scenario where multiple heterogeneous power constraints are practically more relevant. Based on this duality, an algorithm with low complexity is developed. For the more general case of a multicast MBP, a dual problem, which can be solved efficiently, does not exists. However, this thesis presents an equivalent quasi-convex form of the multicast MBP for a special case of long-term CSI which is practically relevant. In addition to the quasi-convex form, a dual problem is also presented. While strong duality is not given in general, a duality-based algorithm finds near optimal solutions with better performance than conventional solutions based on convex relaxation. Based on Multicast Beamforming, this work finally presented an adaptive approach for the global adjustment of the cell pattern of a multicellular network using globally available long-term channel statistics of users. Additionally, users in the network are selected based on instantaneous but local channel statistics. A problem of unicast max-min beamforming (MBF) is a decreased sum rate performance in some cases. One reason is interference due to beam collisions. A formulation of an optimization problem to avoid these collisions is derived and its NP-hardness is proved. Furthermore, this thesis presents several methods with polynomial complexity to find near optimal solutions for this problem. Another reason for an impairment of the sum rate in MMF systems are strongly shadowed users. To improve the sum rate of the MMF system, this thesis presents a solution based on one-way half-duplex relays
- …
