Search

The sparse coding method has been successfully applied to multi-frame super-resolution in recent years. In this paper, we propose a new multi-frame super-resolution framework which combines low-rank fusion with sparse coding to improve the performance of multi-frame super-resolution. The proposed method gets the high-resolution image by a three-stage process. First, a fused low-resolution image is obtained from multi-frame image by the method of registration and low-rank fusion. Then, we use the jointly training method to train a pair of learning dictionaries which have good adaptive ability. Finally, we use the learning dictionaries combined with sparse coding theory to realize super-resolution reconstruction of the fused low-resolution image. As the experiment results show, this method can recover the lost high frequency information, and has good robustness.

The device-to-device (D2D) communication has been regarded as an effective technique for complementing and enhancing the conventional cellular systems owing to its capability of substantially improving both the spectral and power efficiencies of wireless networks. However, the severe interference imposed on the conventional cellular users (CUs) by the geographically close-by D2D pairs may cause a significant performance erosion in the D2D-aided underlaying cellular networks (CNs). In this paper, performance analysis for the D2D-aided underlaying CNs in terms of throughput is provided. We first derive the closed-form expressions of the coverage probability for both the conventional cellular links and the D2D links, followed by giving out the approximated expressions of the ergodic data rate for both an individual cellular/D2D link and the whole underlaying network. Furthermore, the key parameters (e.g., the density of D2D users (DUs) or CUs, and the average geographical distance between a pair of D2D peers) significantly impacting the channel capacity are adaptively adjusted for maximizing the sum data rate of the proposed underlaying networks. In addition, both theoretical analysis and simulation results reveal the attainability of the maximal throughput by optimizing the critical parameters, such as the density of DUs, provided that the scale factor between the DUs and sum users (i.e., comprising both conventional CUs and DUs) can be effectively balanced subject to the constraints specified in the proposed scheme.

Device-to-device (D2D) communication has been a significant supplement for traditional cellular network (CN) when meeting with the increasing number of requiring users. In this paper, we assume that cellular user equipments (CUEs) has priority over D2D user equipments (DUEs) on utilizing spectrum resource, and we aim at maximizing the average number of successfully corresponding DUEs. The spatial distribution of requiring users is assumed as homogeneous Poisson Point Process (PPP), therefore it is analytically tractable to derive the expression of D2D communication coverage probability. We derive the approximate closed-form expression of successfully corresponding DUEs' average number, and then obtain the optimal point of the expression with respect to activated DUE's density, and finally verify the effectiveness of this optimal point by numerical results.

Resource allocation in device-to-device (D2D) aided cellular systems, in which the proximity users are allowed to communicate directly with each other without relying on the intervention of base stations (BSs), is investigated. A new uplink resource allocation policy is proposed for enabling the D2D user equipments (DUEs) to reuse the licensed spectrum, provided that the minimum signal-to-interference (SIR) requirement of conventional cellular user equipments (CUEs) is satisfied. Furthermore, the proposed resource-allocation problem can be formulated as “maximizing the number of simultaneously activated D2D pairs subject to the SIR constraints at both CUEs and DUEs”. Numerical results relying on system-level simulation show that the proposed scheme is capable of substantially improving both the D2D-access probability and the network throughput without sacrificing the performance of conventional CUEs.

In this paper, the resource allocation issues in Device-to-Device (D2D) aided heterogeneous networks are investigated, with successive interference cancellation (SIC) capabilities enabled in base stations (BSs). The problem of maximizing the sum throughput of whole network by considering real-time transmissions is formulated. After that, the globally optimal power control is carried out relying on both the enumeration and Kuhn-Munkres methods. To both guarantee the successful transmission probability (STP) of cellular users (CUEs) and reduce the system’s management overhead, a range-difference based resource allocation scheme is proposed. Numerical results show that the average throughput of CUEs can be substantially improved without sacrificing the performance of CUEs by invoking the proposed resource allocation scheme.

The device-to-device (D2D) communication has been regarded as a promising technique to effectively upgrade the existing cellular network. Despite the intriguing perspectives of D2D technique, the performance of D2D-aided cellular network may degrade owing to the severe interference imposed by newly introduced D2D links. With a motive to deal with this problem, the key performance parameters of D2D-aided underlaying cellular networks, including the coverage probability and total data rate have been investigated. Firstly, we analyze the expressions of coverage probability for both the conventional cellular links and the D2D links, and then we give out the approximate expression of the ergodic rate for both individual links and that of the whole underlaying system. After that, in order to optimize the performance of the system in terms of throughput, more parameters closely related to the channel capacity are studied. Finally, the simulation results revealed that the best values for key parameters (e.g. the density of D2D users) are attainable, and the total data rate can been greatly improved according to our proposed strategies.

Image super-resolution based on learning dictionary has recently attracted enormous interests in the field of image super-resolution. In the general, learning dictionary is trained from a large number of training samples. In this paper, we proposed to use the nonlocal self-similarity to filter blocks that are not similar in structure. Based on the judgment method of structural dissimilarity, we can get a small set of dissimilar training samples and then obtain a pair of high quality Structural Dissimilarity Learning Dictionary (SDLD). Extensive experiments on image super-resolution (SR) are carried out to validate the generality, effectiveness and robustness of the proposed method. Extensive experiments on image super-resolution are carried out to validate the generality, effectiveness and robustness of the proposed method. Experimental results demonstrate that our method can not only reduce the training time largely, but also achieve better results of SR.

In this paper, throughput improvement in device-to-device (D2D) enabled cellular systems is investigated. An uplink resource allocation policy can be implemented, by formulating the reuse relationship as relative position relationship under the constrains of both interference limitation area (ILA) and signal-to-interference-plus-noise ratio (SINR), to find the suitable reuse location relationship between D2D user equipments (DUEs) and cellular user equipments (CUEs). The DUEs can thus be allowed to reuse the licensed spectrum. Furthermore, the proposed resource-allocation algorithm can be readily formulated as an objective function of "maximizing the sum throughput subject to the SINR constrains at both CUEs and DUEs". Numerical results indicate that the proposed algorithm can effectively enhance both the sum throughput and the D2D-access probability (DAP) without seriously impairing the performance of the conventional CUEs.

In this paper, we propose a priority based joint power control and resource allocation algorithm for enhancing the energy efficiency (EE) of device-to-device (D2D) links through enabling the successive interference cancellation (SIC) capabilities at the base stations (BSs). In the power control functional module, the EE-maximization model corresponding to each individual D2D link can be transformed into its equivalent model, in which the objective function is demonstrated to be a concave function of the transmitting power of D2D transmitter and can thus be optimized by implementing joint Bisection and Dinkelbach iterative solution. Inspired by the thought of greedy algorithm, in the radio resource allocation functional module, we propose a priority based resource allocation approach for further reducing the computational complexity of our algorithm. Numerical results show that the proposed approach can substantially improve the EE of the D2D links with rapid converge speed.

This paper proposes a new multiple sparse dictionary model, and realizes super-resolution reconstruction via the multiple sparse dictionary combined with sparse coding, which can improve the resolution of image. The training sample set is constructed by image patches wavelet transform to extract the horizontal, vertical, diagonal and high-frequency information of approximate part composition column feature vectors. Then, calculating a pair of learning dictionary via multiple sparse dictionary which trained by the sparse K-singular value decomposition algorithm. Finally, combining the learning dictionaries and sparse coding theory to realize the image super-resolution reconstruction. As the experiment results show, the new method has good performance for recovering the lost high-frequency information, and has good robustness.

In this paper, the capacity of underlaying cellular networks is analyzed by considering the impact of Device-to-Device (D2D) mode, which is shown to be capable of substantially enhancing the sum transmission rate of the cellular network. Based on the theory of stochastic geometry, the successful transmission probabilities of both the D2D users and the conventional cellular users(CUs) are derived, with the D2D transmission rate as well as the sum rate given out simultaneously. Furthermore, the optimal D2D-users(DUs) density that maximizes the sum rate is also derived. In addition, we extend the analysis from single-cell scenario to the multicell environment. Numerical results demonstrate the relationship between D2D-contributed capacity and the D2D-users density. It is shown that the network capacity can be enhanced by increasing the density D2D users when the D2D-users density is relatively low, until a break-even point is approach, in which case further increasing D2D users may even erode the sum rate.