Spectrum occupancy prediction is a key enabler of agile and proactive decision-making for dynamic spectrum management. In this paper, state-of-the- art statistical models and machine learning prediction methods are evaluated on real-world occupancy time series measured in the Land Mobile Radio bands. While there is no universally best method for forecasting spectrum usage data, significant accuracy improvements are shown to be achievable by selecting a suitable prediction method for different frequencies. Motivated by this observation, we treat the problem of automating the selection of a suitable prediction method from a candidate pool as a machine learning task. An approach is proposed that recommends the best performing method for new data instances by learning from prior predictions on spectrum data with similar characteristics. The merit of this approach is shown in terms of improving the prediction accuracy compared to baseline selections without the recommender.

In this paper, measurements from a spectrum monitoring system are used to investigate sharing scenarios in Land Mobile Radio (LMR) bands. Specifically, channel characteristics which demonstrate sharing opportunities in the time domain in the 450 and 800 MHz LMR bands are presented. Based on the spectrum data, network simulations are conducted to evaluate system performance impacts when the sharing opportunities are exploited to improve spectrum utilization. The results provide improved understanding of appropriate spectrum use models to use for capturing sharing performance in LMR scenarios. It is shown that intelligent spectrum access mechanisms are needed to achieve the best performance trade-offs among utilization efficiency, access delay and airtime.

In this paper, we adopt a multiobjective optimization for the bit and power allocation problem in order to meet the requirements of emerging wireless systems, i.e., achieving higher throughput without considerably increasing the transmit power. More specifically, we propose to simultaneously maximize the throughput and minimize the transmit power of an OFDM system subject to average bit error rate (BER), power budget, and maximum allocated number of bits per subcarrier constraints. The formulated optimization problem is not convex and we use an evolutionary algorithm, i.e., genetic algorithm, in order to obtain the solution. We study the structure of the problem and the obtained solution and notice that the constraint on the average BER can be replaced by a BER per subcarrier constraint. As such, we propose an approximate non-convex optimization problem. We further exploit the structure of the approximate optimization problem and notice that the BER constraint per subcarrier (i.e., the source of the non-convexity) must be satisfied with an equality sign and can be substituted; this leads to an equivalent convex optimization problem where the global optimality of the Pareto solutions is guaranteed. Closed-form expressions are obtained for the bit and power allocations with reduced complexity. Simulation results show that the proposed multiobjective optimization approach provides significant performance improvements over single objective optimization techniques presented in the literature, without incurring additional complexity.

In cognitive radio (CR) networks, secondary users (SUs) are allowed to opportunistically access the primary users (PUs) spectrum to improve the spectrum utilization; however, this increases the interference levels at the PUs. In this paper, we consider an orthogonal frequency division multiplexing OFDM-based CR network and investigate the tradeoff between increasing the SU transmission rate (hence improving the spectrum utilization) and reducing the interference levels at the PUs. We formulate a new multiobjective optimization (MOOP) problem that jointly maximizes the SU transmission rate and minimizes its transmit power, while imposing interference thresholds to the PUs. Further, we propose an algorithm to strike a balance between the SU transmission rate and the interference levels to the PUs. The proposed algorithm considers the practical scenario of knowing partial channel state information (CSI) of the links between the SU transmitter and the PUs receivers. Simulation results illustrate the performance of the proposed algorithm and its superiority when compared to the work in the literature.

A joint channel assignment and power allocation algorithm is presented for cognitive wireless networks, where primary and secondary users operate over the same frequency band at the same time. In this study, we take into consideration a constraint on the interference from secondary users towards primary users, as well as a constraint on the QoS of secondary users. The problem is formulated as a constrained utility maximization problem, for which we prove the existence and the uniqueness of the global optimum solution. Since the problem is NP-hard, we provide a novel heuristic algorithm that operates in two phases. Firstly, it admits all the incoming secondary users and allocates the available channels through a distributed and dynamic procedure, taking into account the interference constraint. Secondly, an iterative power control, which considers both constraints, is applied. In each iteration, a user removal algorithm is employed to reduce the number of admitted secondary users until a steady state is reached. The performance of the joint resource allocation algorithm is investigated in terms of the total number of admitted secondary users for several network parameters, such as the minimum QoS requirements and the interference constraint.

This paper adopts a multiobjective optimization (MOOP) approach to investigate the optimal link adaptation problem of orthogonal frequency division multiplexing (OFDM)-based cognitive radio (CR) systems, where secondary users (SUs) can opportunistically access the spectrum of primary users (PUs). For such a scenario, we solve the problem of jointly maximizing the CR system throughput and minimizing its transmit power, subject to constraints on both SU and PUs. The optimization problem imposes predefined interference thresholds for the PUs, guarantees the SU quality of service in terms of a maximum bit-error-rate (BER), and satisfies a transmit power budget and a maximum number of allocated bits per subcarrier. Unlike most of the work in the literature that considers perfect SU spectrum sensing capabilities, the problem formulation takes into account errors due to imperfect sensing of the PUs bands. Closed-form expressions are obtained for the optimal bit and power allocations per SU subcarrier. Simulation results illustrate the performance of the proposed algorithm and demonstrate the superiority of the MOOP approach when compared to single optimization approaches presented in the literature, without additional complexity. Furthermore, results show that the interference thresholds at the PUs receivers can be severely exceeded due to the perfect spectrum sensing assumption or due to partial channel information on links between the SU and the PUs receivers. Additionally, the results show that the performance of the proposed algorithm approaches that of an exhaustive search for the discrete optimal allocations with a significantly reduced computational effort.

In this paper, we adopt a multiobjective optimization approach to jointly optimize the rate and power in OFDM-based cognitive radio (CR) systems. We propose a novel algorithm that jointly maximizes the OFDM-based CR system throughput and minimizes its transmit power, while guaranteeing a target bit error rate per subcarrier and a total transmit power threshold for the secondary user (SU), and restricting both co-channel and adjacent channel interferences to existing primary users (PUs) in a statistical manner. Since the interference constraints are met statistically, the SU transmitter does not require perfect channel-state-information (CSI) feedback from the PUs receivers. Closed-form expressions are derived for bit and power allocations per subcarrier. Simulation results illustrate the performance of the proposed algorithm and compare it to the case of perfect CSI. Further, the results show that the performance of the proposed algorithm approaches that of an exhaustive search for the discrete global optimal allocations with significantly reduced computational complexity.

In this paper, we investigate the resource allocation problem (joint bit and power loading) of secondary users sharing the radio spectrum with primary users in cognitive radio networks. We consider the co-existence scenario where a secondary user is allowed to access the shared spectrum while guaranteeing tolerable interference to primary users. For such a scenario, we formulate and solve an optimization problem that jointly maximizes the secondary user throughput and minimizes its transmit power while satisfying target bit error rate per subcarrier and certain limits of co-channel and adjacent channel interferences to existing primary users. Simulation results are described that illustrate the performance of the proposed algorithm, and show its closeness to that of an exhaustive search for the equivalent discrete formulation.

This paper studies the joint rate and power allocation problem for OFDM-based cognitive radio systems where secondary users (SUs) can opportunistically access the spectrum of primary users (PUs). We propose a novel algorithm that jointly maximizes the OFDM-based cognitive radio system throughput and minimizes its transmit power, while guaranteeing a target bit error rate per subcarrier and restricting both co-channel interference (CCI) and adjacent channel interference (ACI) to existing primary users. Since estimating the instantaneous channel gains on the links between the SU transmitter and the PUs receivers is impractical, we assume only knowledge of the path loss on these links. Closed-form expressions are derived for the close-to-optimal bit and power distributions. Simulation results are described that illustrate the performance of the proposed scheme and show its closeness to that of an exhaustive search for the discrete optimal allocations. Further, the results quantify the violation ratio of both the CCI and ACI constraints at the PUs receivers due to the partial channel information. The effect of adding a fading margin to reduce the violation ratio is also studied.

This paper proposes a novel low complexity joint bit and power suboptimal allocation algorithm for multicarrier systems operating in fading environments. The algorithm jointly maximizes the throughput and minimizes the transmitted power, while guaranteeing a target bit error rate (BER) per subcarrier and meeting a constraint on the total transmit power. Simulation results are described that illustrate the performance of the proposed scheme and demonstrate its superiority when compared to the algorithm in [4] with similar or reduced computational complexity. Furthermore, the results show that the performance of the proposed suboptimal algorithm approaches that of an optimal exhaustive search with significantly lower computational complexity.

In this paper, a novel joint bit and power loading algorithm is proposed for orthogonal frequency division multiplexing (OFDM) systems operating in fading environments. The algorithm jointly maximizes the throughput and minimizes the transmitted power, while guaranteeing a target average bit error rate (BER) and meeting a constraint on the total transmit power. Simulation results are described that illustrate the performance of the proposed scheme and demonstrate its superiority when compared to the algorithm in [1].

In this paper, a novel low complexity bit and power loading algorithm is formulated for orthogonal frequency division multiplexing (OFDM) systems operating in fading environments and in the presence of unknown interference. The proposed non-iterative algorithm jointly maximizes the throughput and minimizes the transmitted power, while guaranteeing a target bit error rate (BER) per subcarrier. Closed-form expressions are derived for the optimal bit and power distributions per subcarrier. The performance of the proposed algorithm is investigated through extensive simulations. A performance comparison with the algorithm shows the superiority of the proposed algorithm with reduced computational effort.

This paper proposes a novel joint bit and power allocation algorithm for multicarrier systems operating in fading environments. The algorithm jointly maximizes the throughput and minimizes the transmitted power, while guaranteeing a target average bit error rate (BER). Simulation results are described and they illustrate the performance of the proposed scheme and demonstrate its superiority with respect to existing schemes.

Cognitive radio (CR) technology represents a promising solution for solving the problem of spectrum scarcity and underutilization in wireless communications. While orthogonal frequency division multiplexing (OFDM) is regarded as a suitable technology for the CR physical layer, it can suffer from out-of-band radiation, which hinders its spectrum sharing capabilities. In this paper we consider a turbo-coded OFDM based CR system, and propose three methods for shaping the spectrum of the CR signal to reduce interference affecting primary users (PUs) that operate in adjacent bands. The proposed methods use multiple choice sequences (MCS) generated at the bit level by employing different pseudo random bit sequences, by using bit interleavers with different depths, or by a combination of both, respectively. In order to reduce interference to adjacent bands, the sequence with the best spectral shape is transmitted along with its index. Performance of the proposed methods is illustrated with numerical results obtained from simulations, and the trade off between a better spectral shape and the reduction in the spectral efficiency of the systems is discussed.

Cognitive radios must be able to form collaborative wireless network clusters in dynamically changing environments to achieve such desired objectives as interference resilience and low communications overhead. In this work, we explore the affinity propagation (AP) message-passing technique to efficiently group nodes in an ad hoc cognitive radio network (CRN). With the proposed approach, nodes exchange local messages with their immediate neighbours until a high quality set of clusterheads and a corresponding cluster structure emerges. The messages are calculated based on measures of similarity between the network nodes, which are selected based on application requirements and the objective of the grouping process. As an initial application, we focus on finding a small dominating set of a CRN with the aim of reducing the number of nodes that participate in key network functions such as resource management and routing table maintenance. To demonstrate the merits of the proposed clustering approach, the AP technique is evaluated on randomly generated open spectrum access network scenarios. The simulation results demonstrate that the proposed technique provides a smaller number of clusters than methods based on approximating a minimum size dominating set of the corresponding ad hoc network graphs.

Accurate and sparse representation of a moderately fast fading channel using bases functions is achievable when both channel and bases bands align. If a mismatch exists, usually a larger number of bases functions is needed to achieve the same accuracy. In this paper, we propose a novel approach for channel estimation based on frames, which preserves sparsity and improves estimation accuracy. Members of the frame are formed by modulating and varying the band-width of discrete prolate spheroidal sequences (DPSS) in order to reflect various scattering scenarios. To achieve the sparsity of the proposed representation, a matching pursuit approach is employed. The estimation accuracy of the scheme is evaluated and compared with the accuracy of a Slepian basis expansion estimator based on DPSS for a variety of mobile channel parameters. The results clearly indicate that for the same number of atoms, a significantly higher estimation accuracy is achievable with the proposed scheme when compared to the DPSS estimator.

The Ricean K-factor is a practical channel quality measure in many wireless communication applications as it exhibits a fast estimation convergence compared with the explicit estimation of performance metrics such as error rates. Recently, it has been shown that estimation of the K-factor can be improved relative to envelope-based detectors through the use of complex (I/Q) channel observations. In this paper, an analysis of the maximum likelihood estimator of the K-factor from I/Q samples is presented, which illustrates the bias of previous estimators. An improved estimator is then proposed, which has superior bias and efficiency for short data records. For mobile applications, a reliable estimator of the Doppler shift of the specular component is incorporated. Simulation results for a range of channel conditions illustrate that the proposed estimator outperforms prior techniques, and provide insight into the record lengths required to achieve a desired estimation accuracy.

Knowledge of the Ricean K factor is important in many communication applications as it reveals the fading severity experienced by a wireless receiver. In this paper, a statistical analysis of the maximum likelihood estimator (MLE) of K from complex (I/Q) Ricean channel samples is derived. Based on our analysis, an improved estimator of the K factor is proposed with superior bias and efficiency properties for short data records. The lack of robustness of the tractable MLE in mobile applications is addressed by incorporating a reliable estimator of the specular Doppler shift. The simulation results show that the proposed technique can outperform prior methods for I/Q mobile channel snapshots.

To enable closed-loop operation of high-mobility systems, reliable channel prediction is needed to prevent the obsolescence of channel information due to feedback and processing delays. In this paper, novel linear pilot symbol-aided techniques are proposed for block-adaptive prediction of flat time-selective Rayleigh fading channels. Unlike many previous investigations, which assume that the fading statistics are known, we focus on practical conditions where prediction must take place over short noisy windows of unknown channels. Significant performance gains are observed when the method is applied, for example, to a selection transmitter diversity system operating in a fast fading environment.

Exploiting the full spatial diversity available in mobile wireless channels is most effective when some information about the channel is available at the transmitter. In many practical applications, such information is rapidly outdated and has limited realizable benefits. This note investigates the feasibility of linear fading prediction, applied to noisy channel estimates. The predictions are then used for antenna subset selection for space-time block coding. It is shown, using synthesized and measured channel data, that multidimensional prediction from short channel snapshots is unreliable for dense scattering channels. However, the use of parallel predictors can provide a significant increase in the diversity, and hence performance, achievable with antenna subset selection relative to outdated or no channel information at the transmitter.

Autoregressive stochastic models for the computer simulation of correlated Rayleigh fading processes are investigated. The unavoidable numerical difficulties inherent in this method are elucidated and a simple heuristic approach is adopted to enable the synthesis of accurately correlated, bandlimited Rayleigh variates. Startup procedures are presented, which allow autoregressive simulators to produce stationary channel gain samples from the first output sample. Performance comparisons are then made with popular fading generation techniques to demonstrate the merits of the approach. The general applicability of the method is demonstrated by examples involving the accurate synthesis of nonisotropic fading channel models.

A new parametric approach is proposed for the estimation of the maximum Doppler frequency, or equivalently the mobile speed, in narrowband mobile radio channels. In this work, simple and efficient Doppler-frequency estimators are formulated based on using a small number of samples of the channel autocorrelation function. Unlike previous approaches, the chosen parameterization is shown to be robust in a microcellular propagation environment, which may be characterized by nonisotropic scattering and/or a specular component of unknown strength. Simulation results are described to illustrate the effects of additive noise and the finite data-record performance.

The computer generation of multiple cross-correlated Rician fading channels is investigated. We prove that the output sequences of existing multichannel fading simulators are restricted to have cross-correlation statistics that have the same functional form as the component autocorrelation functions. To overcome this limitation, vector autoregressive stochastic models are proposed for the generation of multiple Rician fading processes with specified realizable autocorrelation and cross-correlation statistics. This capability is desirable, for example, to permit realistic performance assessments of space-time modem designs by enabling the simulation of space-time-selective wireless channel models. The utility of the simulation approach is demonstrated by the accurate synthesis of some bandlimited multichannel Rayleigh and Rician processes.

Accurate estimation of the Doppler spread of a mobile radio channel is of significant value as it reveals the channel's rate of change. This information can then be used to improve the performance of many wireless communication subsystems. In this paper, a nonparametric technique is proposed to estimate the maximum Doppler frequency of narrow band fading channels. Moreover, a computationally efficient periodogram-based approach is formulated, and its finite sample performance is assessed and compared to parametric covariance-based estimators via computer simulations. On the basis of these comparisons, we claim that the simplicity of the proposed method, its accurate performance for small sample sizes in a wide range of propagation scenarios, and its robustness to low SNR conditions make it a strong competitor to the parametric approaches.

Vector autoregressive stochastic models are proposed for the synthesis of multiple Rayleigh fading processes with specified autocorrelation and cross-correlation functions. This capability is desirable, for example, to permit realistic performance assessments of space-time modem designs by enabling the simulation of space-time-selective mobile radio channel models. The utility of the simulation approach is demonstrated by the accurate generation of some bandlimited vector Rayleigh processes.

Autoregressive stochastic models for the computer simulation of correlated Rayleigh fading processes are investigated. The unavoidable numerical difficulties faced by this method are elucidated and a simple heuristic approach is adopted to enable the synthesis of accurately correlated, bandlimited Rayleigh variates. Performance comparisons are made with a popular generation technique to demonstrate the merits of the method.