Due to limited cooperation between the primary users and the secondary users (SUs) in practical spectrum sharing systems, the secondary transmitters and receivers are assumed to have partial channel state information related to the primary receiver. Under such an assumption, this work investigates power allocation strategies for the SUs subject to an outage probability constraint on the primary transmission and a peak transmit power constraint on the secondary transmission. The challenge lies in the non-convexity of the outage probability constraint. Firstly, we prove that strong duality holds and that the Karush-Kuhn-Tucker (KKT) conditions are necessary for optimality. The optimal solution is then derived by addressing the optimality issues of the KKT solutions. Secondly, in order to further reduce the algorithmic complexity, two suboptimal strategies are proposed. The first one is designed based on several simplifications of the optimal strategy. The second one is derived from the convex relaxation of the non-convex primal problem, which corresponds to the problem with the conventional interference temperature constraint. The performance for both suboptimal strategies is derived in closed form. All proposed strategies are shown to outperform non-adaptive power transmission. The near-optimality of the two suboptimal strategies is also validated, in particular for the first one.

In wireless communication systems, dual-polarized (DP) instead of single-polarized (SP) multiple-input multiple-output (MIMO) transmission is used to improve the spectral efficiency under certain conditions on the channel and the signal-to-noise ratio (SNR). In order to identify these conditions, we first propose a novel channel model for DP mobile Ricean MIMO channels for which statistical channel parameters are readily obtained from a moment-based channel decomposition. Second, we derive an approximation of the mutual information (MI), which can be expressed as a function of those statistical channel parameters. Based on this approximation, we characterize the required SNR for a DP MIMO system to outperform an SP MIMO system in terms of the MI. Finally, we apply our results to channel measurements at 2.53 GHz. We find that, using the proposed channel decomposition and the approximation of the MI, we are able to reproduce the SNR values above which DP MIMO systems outperform SP MIMO systems.

This paper investigates power allocation strategies for secondary users (SUs) in cross-interfering spectrum sharing systems. Addressing limited cooperation between primary users (PUs) and SUs, only instantaneous channel state information (CSi) of the secondary link and statistical CS! of the other links is assumed to be available at the secondary transmitters (STs). We aim at maximizing the secondary achievable rate subject to both a peak power constraint at the ST and an average interference power constraint at the primary receiver. First, the optimal power control strategy is developed. !n order to reduce the complexity, two suboptimal optimization strategies are proposed. The first one, named double threshold waterfilling (DT-WF), is based on an approximation of the optimal solution. The second strategy, named double threshold constant-power waterfilling (DTCP-WF), further simplifies DT-WF. Additionally, the achievable performance is derived in closed form for both suboptimal strategies. We also discuss the algorithm design in the multiple primary links scenario. Numerical results show the effectiveness of the proposed strategies and validate the accuracy of the closed-form analysis.

Due to limited cooperation between the primary users and the secondary users (SUs), the SUs only have partial channel state information related to the primary receiver. Under such an assumption, this paper investigates power allocation strategies for the SUs in spectrum sharing systems subject to an outage probability constraint on the primary transmission and a peak transmit power constraint on the secondary transmission. The challenge lies in the non-convexity of the outage probability constraint. We investigate two methods to solve the non-convex optimization problem. First, strong duality is shown to be satisfied. Since the Karush-Kuhn-Tucker (KKT) conditions are then necessary but not sufficient for optimality, we derive the KKT solutions and address the optimality issue of the solutions. Alternatively, the non-convex constraint is replaced by a stricter convex constraint. A more conservative suboptimal solution is thus obtained, which corresponds to the solution of the problem with the conventional interference temperature constraint. Simulation results show that both proposed strategies outperform constant-power transmission. The superiority of the optimal strategy over the suboptimal one is also verified.

Dual-polarized (DP) multiple-input multiple-output (MIMO) systems are known to be beneficial in terms of the spectral efficiency under certain conditions on the channel and the signal-to-noise ratio (SNR). In order to assess under which conditions DP antenna setups are advantageous over single-polarized (SP) antenna setups, an accurate yet analytically tractable modeling of such channels is a necessary but challenging task. We propose a novel channel model for DP mobile Ricean MIMO channels for which the statistical channel parameters are readily obtained from a moment-based channel decomposition. An approximate evaluation of the mutual information (MI), expressed as a function of the statistical channel parameters of the proposed channel model, is derived. This approximation allows to assess when it is beneficial in terms of the MI to switch from an SP to a DP setup. Finally, we evaluate the MI for selected SP and DP setups based on channel measurements at 2.53 GHz. We find that the DP setup only provides an improvement in terms of the MI in medium- to high-K-factor scenarios above a certain SNR. The approximate evaluation of the MI is able to accurately reproduce the SNR at which one should switch from an SP to the DP setup.

We propose a general channel model for mobile Ricean multiple-input multiple-output (MIMO) channels, which is characterized by parameters that are readily obtainable from measurements. To this end, a moment-based channel decomposition is derived. For the case of statistical channel state information (CSI) at the transmitter (TX) and instantaneous CSI at the receiver, we derive an approximation of the achievable rate, i.e., the mutual information (MI), which is only a function of the channel parameters of the proposed channel model and thus gives insight into the parameters that influence the MI. Finally, we evaluate the MI for a 4 × 4 MIMO system based on channel measurements at 2.53 GHz for two low-complexity TX strategies, i.e., a beamforming (BF) and a spatial multiplexing (SM) strategy. We find that the proposed channel model and the approximate evaluation of the MI are both accurate for realistic signal-to-noise ratio (SNR) values. The approximate evaluation is able to reproduce crossing points between the MI of the considered TX strategies; it thus reflects the SNR at which one should switch from the BF to the SM strategy.

The extension of multiple-input multiple-output (MIMO) systems to the polarization domain is known to have potential benefits in terms of the spectral efficiency. It is thus of importance to understand how the channel influences the spectral efficiency, e.g., when a dual-polarized (DP) antenna setup should be favored over a single-polarized (SP) one. In this work, we study the achievable rate, i.e., the sum mutual information (MI) of all transmitted streams, over DP MIMO channels with linear receivers. We derive an approximation of the sum MI with a linear minimum mean squared error (LMMSE) or a zero forcing receiver which is an explicit function of statistical channel parameters. The evaluations of selected 4×4 SP and DP MIMO setups are based on channel measurements performed at 2.53 GHz in an urban macrocell scenario. The results reveal that DP antenna setups can yield substantial gains in the sum MI on links with high K-factors and high signal-to-noise ratios (SNRs). Moreover, the degradation in the sum MI with two streams due to the use of an LMMSE receiver is found to be small, especially for the DP antenna setup. The approximate evaluation of the sum MI is able to reproduce the SNR threshold region in which a second transmitted stream should be activated or one should switch from an SP to a DP antenna setup.

This paper studies an optimal power allocation strategy in spectrum-sharing cognitive radio systems. In reality, it is difficult for the secondary users (SUs) to obtain perfect channel state information (CSI) related to the primary users (PUs) due to the lack of cooperation between the SUs and the PUs. We assume that only instantaneous CSI of the secondary link and statistical CSI of the other links is available to the SUs. This assumption not only reveals the realistic challenges, but also brings mathematical challenges in solving the optimization problem in closed form, since the performance evaluated with statistical CSI involves expectation operations. We reformulate the objective function in closed form and derive the optimal power allocation solution. The feasibility of the solution is discussed. Furthermore, we provide several interesting insights into the achievable performance based on the properties of the solution.

In many typical mobile communication receivers, the channel is estimated based on pilot symbols to allow for a coherent detection and decoding in a separate processing step. Currently, much work is spent on receivers which break up this separation, e.g., by enhancing channel estimation based on reliability information on the data symbols. In this paper, we evaluate the possible gain of a joint processing of data and pilot symbols in comparison to the case of a separate processing in the context of stationary Rayleigh flat-fading channels. Therefore, we discuss the nature of the possible gain of a joint processing of pilot and data symbols. We show that the additional information that can be gained by a joint processing is captured in the temporal correlation of the channel estimation error of the solely pilot-based channel estimation, which is not retrieved by the channel decoder in case of separate processing. In addition, we derive a new lower bound on the achievable rate for joint processing of pilot and data symbols. Finally, the results are extended to multiple-input multiple-output channels.

Interference is one of the major bottlenecks in current cellular networks. A realistic evaluation of the achievable performance in interference-limited systems based on measured channels is thus necessary; however, only few results are known from literature. We evaluate the achievable performance in a cellular network with inter-cell interference based on measured channels in an urban macrocell scenario at 2.53 GHz. To this end, the mutual information is introduced as an appropriate performance measure over fast and slow fading channels that are non-stationary but doubly underspread. We discuss appropriate channel normalizations and the limitations of an evaluation based on sequential measurements. Moreover, we analyze the accuracy of an approximate but commonly used evaluation of the mutual information. A second-order multivariate Taylor series expansion reveals the signal and interference contributions to the approximation error. We find that the approximation is accurate in realistic ICI scenarios.

A common simplification in the treatment of random linear channels is the wide-sense stationary and uncorrelated scattering (WSSUS) assumption. For wireless channels, this assumption is, however, only fulfilled in an approximative sense inside local time-frequency regions. Since algorithms in wireless digital communications often rely on knowledge of second order statistics of the channel, it is important to know the size of local quasi-stationarity regions. Thus, we determine quasi-stationarity regions in distance for an urban macrocell scenario. We observe that, based on the chosen measure and in our specific scenario, the time-frequency properties are dominant compared to the spatial properties in defining the size of quasi-stationarity regions. Furthermore, we find that in some cases the quasi-stationarity regions strongly depend on the mobile terminal orientation.

A common simplification in the treatment of random linear channels is the assumption of stationarity of the channel in time. The wireless channel is, however, known to be inherently non-stationary. We detail a methodology for non-stationarity analysis from an algorithmic perspective. For the determination of local quasi-stationarity regions, we consider a multi-link downlink scenario where multiple base stations use transmit beamforming to concurrently transmit to a single mobile terminal per time slot. We obtain an algorithm-specific measure, i.e., the signal-to-interference-plus-noise ratio (SINR) degradation, with which we evaluate local quasi-stationarity regions. Furthermore, we relate and compare the SINR degradation to the correlation matrix distance (CMD). In an urban macrocell scenario relevant to 3GPP Long Term Evolution (LTE), we find that the resulting local quasi-stationarity distances show the same trends, but that the CMD overestimates the average distances for our system model.