Semisupervised band selection (BS) technique plays an important role in processing hyperspectral images (HSIs) because of its superiority of using the limited labeled data and plentiful unlabeled data to select the discriminative and informative feature, which copes with the problem of high-dimensional and scare labeled samples of HSIs. Among semisupervised BS models, graph-based methods are superior to others in many situations and have received more and more attention. However, traditional graph-based models construct a similarity matrix and select valuable bands independently. The similarity matrix remains constant, which will damage the local manifold structure and lead to a suboptimal result. To solve this problem, we propose a semisupervised band selection with an optimal graph (BSOG) approach, which performs BS and local structure learning simultaneously. Instead of fixing the input similarity matrix, the similarity matrix is updated constantly to learn a better local structure. Besides, the learned similarity matrix is adaptive. Then, the optimal band subset can be selected by analyzing the obtained projection matrix$\boldsymbol W$. An efficient and simple optimization algorithm is proposed to solve this model. Experiments on four real HSIs data validate the effectiveness of the proposed model.

Recently, with the explosive increase in Internet data, the traditional Graph-based Semi-Supervised Learning (GSSL) model is not suitable to deal with large scale data as the high computation complexity. Besides, GSSL models perform classification on a fixed input data graph. The quality of initialized graph has a great effect on the classification result. To solve this problem, in this paper, we propose a novel approach, named optimal bipartite graph-based SSL (OBGSSL). Instead of fixing the input data graph, we learn a new bipartite graph to make the result more robust. Based on the learned bipartite graph, the labels of the original data and anchors can be calculated simultaneously, which solves co-classification problem in SSL. Then, we use the label of anchor to handle out-of-sample problem, which preserves well classification performance and saves much time. The computational complexity of OBGSSL is O(ndmt+nm2), which is a significant improvement compared with traditional GSSL methods that need O(n2d+n3), where n, d, m and t are the number of samples, features anchors and iterations, respectively. Experimental results demonstrate the effectiveness and efficiency of our OBGSSL model.

As the captured information in our real word is very scare and labeling sample is time cost and expensive, semisupervised learning (SSL) has an important application in computer vision and machine learning. Among SSL approaches, a graph-based SSL (GSSL) model has recently attracted much attention for high accuracy. However, for most traditional GSSL methods, the large-scale data bring higher computational complexity, which acquires a better computing platform. In order to dispose of these issues, we propose a novel approach, bipartite GSSL normalized (BGSSL-normalized) method, in this paper. This method consists of three parts. First, the bipartite graph between the original data and the anchor points is constructed, which is parameter-insensitive, scale-invariant, naturally sparse, and simple operation. Then, the label of the original data and anchors can be inferred through the graph. Besides, we extend our algorithm to handle out-of-sample for large-scale data by the inferred label of anchors, which not only retains good classification result but also saves a large amount of time. The computational complexity of BGSSL-normalized can be reduced to O(ndm +nm2), which is a significant improvement compared with traditional GSSL methods that need O(n2d + n3), where n, d, and m are the number of samples, features, and anchors, respectively. The experimental results on several publicly available data sets demonstrate that our approaches can achieve better classification accuracy with less time costs.

Unmanned aerial vehicle (UAV)-satellite communication has drawn dramatic attention for its potential to build the integrated space-air-ground network and the seamless wide-area coverage. A key challenge to UAV-satellite communication is its unstable beam pointing due to the UAV navigation, which is a typical SatCom on-the-move scenario. In this paper, we propose a blind beam tracking approach for Ka-band UAV-satellite communication system, where UAV is equipped with a hybrid large-scale antenna array. The effects of UAV navigation are firstly released through the mechanical adjustment, which could approximately point the beam towards the target satellite through beam stabilization and dynamic isolation. Specially, the attitude information for mechanical adjustment can be realtimely derived from data fusion of low-cost sensors. Then, the precision of beam pointing is blindly refined through electrically adjusting the weight of the massive antennas, where an array structure based simultaneous perturbation algorithm is designed. Simulation results are provided to demonstrate the superiority of the proposed method over the existing ones.

In this paper, we design a channel tracking method for massive multiple-input multiple-output systems under both time-varying and spatial-varying circumstances. By exploiting the characteristics of massive antenna array, a spatial-temporal basis expansion model is proposed to reduce the effective dimension of uplink/downlink channel, which decomposes channel state information into time-varying spatial information and gain information. We first model the user's movement as the one-order unknown Markov process, whose parameters are blindly obtained by expectation and maximization learning. Then, the uplink time-varying spatial information can also be blindly tracked by unscented Kalman filter and Taylor series expansion of the steering vector, while the rest of uplink channel gain information can be trained by only a few pilot symbols. Due to physical angle reciprocity, the spatial information of the downlink channel can be immediately computed from the uplink counterpart, which greatly reduces the complexity of downlink channel tracking. Various numerical results are provided to demonstrate the effectiveness of the proposed method.

mmWave massive multiple-input multiple-output (MIMO) system has gained much attention for its considerable improvement in system throughput. However, the cost of the complex hardware, e.g., the radio frequency (RF) chains, hinders it from the practical deployment. In this paper, we propose an angle space channel tracking method for mmWave massive MIMO systems with limited RF chains (hybrid scheme). Specifically, the users can be scheduled according to their DOA information, i.e. angle division multiple access (ADMA). Besides, the channel information can be divided into direction of arrival (DOA) information and gain information respectively, where DOA can be tracked through unscented Kalman filter (UKF), while the gain information can be obtained from beam training and spatial rotation. Numerical results are provided to corroborate our studies.

In this paper, we propose a new channel tracking method for massive multiple-input multiple-output (MIMO) systems under both the time-varying and spatial-varying circumstance. With spatial-temporal basis expansion model (ST-BEM), the channel information is decomposed into the spatial information and gain information, where the former is determined by the central angle as well as the angular spread of the incoming signal. We first blindly track the central angle by the extended Kalman filter (EKF) and obtain the angular spread through Taylor series expansion of the steering vector. Then, the channel gain information can be estimated with only a few pilot symbols. Various numerical results are provided to demonstrate the effectiveness of the proposed method.

The millimeter-wave (mm-wave) massive multiple-input multiple-output (MIMO) system has gained much attention for its considerable improvement in system throughput. However, the cost of complex hardware, e.g., radio frequency (RF) chains, hinders it from practical deployment. In this paper, we propose an angle domain hybrid precoding and channel tracking method by exploring the spatial features of the mm-wave massive MIMO channel. The number of the effective spatial beams, or equivalently the RF chains, is enormously decreased via the operation of spatial rotation. The users are then scheduled by the angle division multiple access scheme, which groups users according to their direction of arrivals (DOAs). Meanwhile, a channel tracking method is designed for the subsequent data transmission through a small number of pilot symbols. Specifically, the channel information is divided into the DOA information and the gain information, where the DOA information is tracked by a modified unscented Kalman filter and the gain information is estimated from beam training. Numerical results are provided to corroborate our studies.

An improved quadrature nonlinear filter named delayed Genz-Keister sequences-based sparse-grid quadrature filter (DGKSGQF) is developed for the target tracking problems. The filter changes the non-nested Gaussian quadrature points of the quadrature filters to the nested Genz-Keister points for selecting the unvariate points, which are the basis point sets extended to form a multidimensional grid using the sparse-grid theory. As a result, the points used for lower accuracy levels DGKSGQF can be reused for any higher accuracy level. Thus, it can further reduce the number of total points used for the conventional Gauss-Hermite SGQF without sacrificing performance. The proposed filter is applied to the reentry ballistic target tracking problem. The simulation results show that the DGKSGQF achieves higher accuracy than the EKF and the UKF. In addition, it can more flexibly control the performance in terms of the number of points and accuracy level.

This paper presents a method using the Global Positioning System (GPS)-measured course angle to improve the accuracy of the vehicle attitude estimation for low-cost inertial navigation system/GPS (INS/GPS) integration. Observability properties of the error states in the low-cost integration navigation system are first analyzed, indicating that the attitude estimation is severely affected by vehicle maneuvers, particularly the yaw angle. The pitch and roll angles are strongly observed; hence, the observability of these two angles is nearly free of influence caused by vehicle maneuvers, and these two angles can be accurately estimated. To improve the yaw-angle estimation, we propose a cascaded Kalman filter to deal with the yaw angle separately with the aid of the GPS-measured course angle. Additionally, two switching rules are established to remove the influence caused by the sideslip angle and GPS noise. The experimental results validate the observability analysis of the low-cost INS/GPS system and show that the proposed attitude estimation method can effectively improve the accuracy of the vehicle attitude estimation, suggesting that this technique is a viable candidate for many control applications used in cars.