An improved blind spectrum sensing scheme is established by the covariance matrix Cholesky decomposition and radial basis function (RBF)-support vector machine (SVM) decision classification at low signal-to-noise ratios (SNRs). Under strong background noises, the proposed scheme improves the recognition rate of primary users (PUs) than that of the current blind spectrum sensing. First, the ratio of the maximum-to-minimum eigenvalue of a covariance matrix obtained by the Cholesky decomposition is used to construct the statistics. Second, the statistics are labeled with “+1” or “-1,” namely, the energy characteristics of the training samples are extracted and marked with “+1” for PUs and “-1” for noises. Finally, an RBF-SVM classification model, with an intelligent RBF as the SVM kernel function, is obtained by training the above-mentioned statistics and the labels. Thus, the received signals are classified as PUs or not be trained in the SVM model. The threshold possesses self-learning ability, and it distinguishes PU signals from noises effectively. The classification among PU signals and noises is implemented by the optimal SVM decision boundary, derived from maximizing the margin of the decision boundary of trained samples for efficient detection. In addition, the complexity of the statistic construction is lower than that of the conventional maximum minimum eigenvalue (MME). The simulation results show that the RBF in our scheme has 77.5% accuracy at -10 dB, and it outperforms linear kernel function significantly by about 27.5% in accuracy at -10 dB. In addition, the average error probability of the proposed scheme is reduced by about 26% when compared with those of original SVM schemes at -20 dB. The proposed scheme also outperforms the current MME detection in detection probability over 10% at -20 dB. Therefore, the proposed blind spectrum sensing scheme can be efficiently used to detect the PUs by the covariance matrix Cholesky decomposition and the RBF-SVM decision classification in the fifth-generation (5G) communications, especially at low SNRs.

An optimized design of a sparse code multiple access (SCMA) codebook for uplink wireless communications is presented by dividing an optimized 16-point round quadrature amplitude modulation (QAM) into several subsets. The main goal of the scheme is to maximize the minimum Euclidean distance (MED) and thus reduces the collisions of the information bits on the resources. The final SCMA codebook is obtained with the mapping matrix, which indicates the sub-constellations generating by dividing the mother QAM constellation. Simulation results show that, in a Nakagami fading channel, the optimized SCMA scheme by the proposed design method achieves significantly performance gains approximately 1.0, 1.4, 2.5, 3.5, and 4.0 dB at bit error rate (BER) of 10–4, respectively, when compared with those of an undivided 16-QAM constellation, a trellis code modulation (TCM) division, an original SCMA codebooks, a low-density signature (LDS), and an irregular LDS (IrLDS) schemes. In addition, at the signal-to-noise ratio (SNR) ranging from 0 to 10 dB, the constellation constrained capacity of the scheme by the proposed method achieves more gains over those of the original SCMA, TCM, the undivided star-QAM, the LDS, and the IrLDS schemes. Thus, it can be combined with the grant-free random access mechanism to obtain rapid and low-cost access in next-generation wireless packet services and other applications.

Unmanned aerial vehicles (UAVs) have gained much popularity for a wide variety of applications that do not require human operators, but also face the spectrum scarcity problem. To address this problem, a cognitive UAV network (CUAVN) framework based on cognitive radio (CR) technology is formulated in this article. On the basis of this, the sequential probability ratio test (SPRT) is used as a fundamental cooperative spectrum sensing (CSS) scheme to significantly reduce the average number of samples, but the uncertainty of its detection delay has a negative impact on the spectrum-sensing performance and the achievable throughput of CUAVNs, especially in the multiband spectrum sensing. Considering the location flexibility and uneven distribution density of UAVs, we design an intraframe CSS structure, with aim of achieving CSS between microsensing slots. Furthermore, a sequential maximal truncation (SMT) is proposed to realize a quick multiband spectrum sensing for delay-constraint CUAVNs. Compared to uniform tail truncation (UTT), simulation results show that the proposed quick multiband spectrum sensing with the help of SMT requires fewer sensing times in support of better sensing performance and achievable throughput, especially under a strict delay constraint.

The spectrum scarcity and low utilization rates constrain further development possibilities in wireless sensor networks, the cognitive radio (CR) enables multiple sensor nodes (SNs) to detect whether the primary node (PU) is occupying spectrum bands by cooperative spectrum sensing (CSS) and provides an opportunity to access unused channels. To address the challenges posed by the inherent characteristics of wireless channels, CSS have emerged in a cognitive wireless sensor network (CWSN). However, this also provides opportunities for malicious sensor nodes (MSNs) to manipulate spectrum sensing data by Byzantine attack. Therefore, in this paper, we employ support vector machine (SVM) to define the optimal hyperplane to distinguish MSNs from normal sensor nodes (NSNs), and combine it with a novel algorithm associated with reputation values to exclude sensing results of MSNs, thereby improving the CSS performance of CWSNs. Furthermore, we also compare this enhanced SVM (ESVM) algorithm with other SVM approaches, evaluating the system’s detection probability, variations with changing parameters and the stability of sensing performance. A series of simulation results demonstrate that ESVM exhibits better robustness against Byzantine attack, and the sensing accuracy is increased by 10.77%-14.21% compared with the existing method in the same scenario.

An enhanced interference alignment strategy with minimum interference leakage (MIL) criterion and Riemannian conjugate gradient (RCG) algorithm is proposed for intelligent reflecting surface (IRS)-assisted multiuser multiple-input multiple-output (MIMO). In this letter, the maximum sum rate is formulated as the optimization objective, with alternate optimization of phase shift vector at IRS as well as precoding and interference suppression vectors at transceivers respectively. MIL interference alignment criterion is used to iteratively solve the precoding vector and the interference suppression vector via channel reciprocity property. RCG algorithm is further applied to derive the IRS phase shift vector and maximize sum rate with the condition of ensuring a given minimum interference leakage threshold to eliminate system interference. Simulation results reveal that the proposed strategy effectively enhances sum rate performance compared with the scheme using random IRS phase shift vector and random precoding/interference suppression vectors as well as ”AP+RCG” scheme in IRS-assisted multiuser single-input single-output (SISO) scenario. In addition, compared with ”MMSE+RCG” strategy in the case of same multi-antenna transceiver pairs and antenna numbers, the proposed strategy could make a tradeoff between sum rate performance and computational complexity.

In current long term evolution (LTE) system, resource efficiency of the random access procedure is low due to the semi-static resource allocation strategy. With the massive access requirement in the future 5G network, this problem would become more and more serious. To effectively reduce the resource consumption while guaranteeing the access delay and reliability, intelligent access channel allocation must be applied, which requires the base station (BS) to accurately estimate the number of user equipments (UEs) that are performing random access. Motivated by this, we propose a novel LSTM-based machine learning approach for active UE number estimation in random access procedure. Meanwhile, a simple minimum distance method is devised as the baseline. Our proposed LSTM-based approach can also predict the UE number at the next random access procedure. Simulation results show that the proposed method achieves quite accurate estimation and prediction of UE number.