Recently, salient object detection (SOD) methods have achieved impressive performance. However, salient regions predicted by existing methods usually contain unsaturated regions and shadows, which limits the model for reliable fine-grained predictions. To address this, we introduce the uncertainty guidance learning approach to SOD, intended to enhance the model's perception of uncertain regions. Specifically, we design a novel Uncertainty Guided Refinement Attention Network (UGRAN), which incorporates three important components, i.e., the Multilevel Interaction Attention (MIA) module, the Scale Spatial-Consistent Attention (SSCA) module, and the Uncertainty Refinement Attention (URA) module. Unlike conventional methods dedicated to enhancing features, the proposed MIA facilitates the interaction and perception of multilevel features, leveraging the complementary characteristics among multilevel features. Then, through the proposed SSCA, the salient information across diverse scales within the aggregated features can be integrated more comprehensively and integrally. In the subsequent steps, we utilize the uncertainty map generated from the saliency prediction map to enhance the model's perception capability of uncertain regions, generating a highly-saturated fine-grained saliency prediction map. Additionally, we devise an adaptive dynamic partition (ADP) mechanism to minimize the computational overhead of the URA module and improve the utilization of uncertainty guidance. Experiments on seven benchmark datasets demonstrate the superiority of the proposed UGRAN over the state-of-the-art methodologies. Codes will be released at https://github.com/I2-Multimedia-Lab/UGRAN.

Multimodal data fusion has a long research history since audio-visual speech recognition, which is inspired by the McGurk effect. Because of the limited model capacity of the traditional methods, multimodal data fusion researches are not so popular for a period. Recently, the advances of deep learning techniques open up new opportunities for the multimodal data fusion field. However, there is still a great gap in multimodal data processing ability between artificial intelligence and human beings. Many problems in multimodal data processing are still necessary to be researched. In this work, we propose to gain an insight into the information fusion level and apply different information fusion strategy to different situations. We analyze the different situations of the multimodal data fusion process and divide them into two categories, including consistent information fusion and contradictory information fusion. We demonstrate some toy examples of the different cases of the multimodal data fusion process.

Load balancing (LB) is important in heterogeneous networks (HetNet). This paper investigates effective cell association schemes for LB. Considering the centralized cellular network architecture with Super Base Stations (SBS), a centralized cell association scheme is proposed, where the calculations are all performed in the SBS and information exchange within the SBS is convenient. There is also no need for the users to feed back information. Thus the signaling overhead is reduced a lot. Moreover, to alleviate the sensitiveness to load changes, the load fairness index (LFI) is defined. The proposed scheme is not performed until the LFI is lower than a given threshold. It is verified by simulations that the proposed centralized cell association scheme can adjust the cell-specific bias according to the cell load and provide better geometry mean rate (GMR) and outage probability (OP) over the comparing schemes. Moreover, with the introduction of LFI, the number of LB performed can be effectively reduced with a lower LFI threshold, however, the system performance such as GMR and OP also degrades. Therefore, an appropriate LFI threshold is needed to balance the system performance and complexity.

This paper presents a novel and practical study on the position-based radio propagation channel for High-Speed Railway by performing extensive measurements at 2.35 GHz in China. The specification on the path loss model is developed. In particular, small scale fading properties such as K-factor, Doppler frequency feature and time delay spread are parameterized, which show dynamic variances depending on the train location and the transceiver separation. Finally, the statistical position-based channel models are firstly established to characterize the High-Speed Railway channel, which significantly promotes the evaluation and verification of wireless communications in relative scenarios.

In this paper, a novel call admission control (CAC) scheme for multicast and unicast integrated services is proposed, which optimizes service provision capacity for both services. A resource sharing model is presented in which multiple classes services are dynamically admitted into the system. Several basic CAC rules for the above resource model are proposed first and improved mechanisms are investigated to further increase system performance. Subsequently, the call level QoS and system performance of the resulting CAC scheme are evaluated. Numeric results show that the proposed scheme can achieve optimal multiclass service performance while satisfying QoS constraints.

With its new features such as multi-homing and multi-streaming the Stream Control Transmission Protocol (SCTP) has become a promising candidate as a general-purpose transport layer protocol. Multi-homing in an SCTP association can make concurrent multi-path transfer an appealing candidate to satisfy the ever increasing user demands for bandwidth. Multiple streams provide an aggregation mechanism to accommodate heterogeneous objects, which belong to the same application but may require different QoS from the network. However, the current approach lacks an internal mechanism to support preferential treatment among its streams for concurrent multi-path transfer. In this paper, we introduce WM2-SCTP (Wireless Multi-path Multi-flow - Stream Control Transmission Protocol), a transport layer solution for concurrent multi-path transfer with parallel subflows. WM2-SCTP aims at exploiting SCTP's multi-homing and multi-streaming capability by grouping SCTP streams into subflows based on their required QoS and selecting best paths for each subflow to improve data transfer rates. The results show that under different scenarios WM2-SCTP, can effectively enhance transmission efficiency.

In this paper, a novel radio admission control (RAC) scheme is proposed for handling multiclass services in Long Term Evolution (LTE) systems. An objective function of maximizing the number of admitted users is proposed to evaluate the system capacity. To solve the optimization problem, we present a combined complete sharing (CS) and virtual partitioning (VP) resource allocation model and develop a service degradation scheme in case of resource limitations in the proposed RAC scheme. Call blocking probability, system resource utilization and system capacity are used as performance metrics and are evaluated by using a K-dimensional Markov Chain model. Numerical results show that an optimal proportion of resource deployment for different service groups can be identified to maximize system capacity while at the same time maintaining quality of service (QoS) constraints of all admitted users.

In order to improve the efficiency of paging management, various paging strategies have been proposed for tracking single mobile user in the wireless cellular network. However, most of the schemes have ignored the important problem of efficient search for multiple mobile users under delay and bandwidth constraints. Given the condition that the search is over only after all the users in the group are found, this problem is defined as the Conference Call Search (CCS) problem. As opposed to the single user tracking, for which one can always reduce the expected search cost, for a multicast search the dependency between the delay and the search cost is NP-hard. In this paper, we propose a non-optimal search method that yields a low search delay as well as a low search cost. The 2D Markov walk is used as the mobility model which describes a broad class of mobility patterns. The Weighted Bipartite Graph Matching is used to allocate paging bandwidth which keeps a total maximum location probability at each paging round. Experimental results show that our paging scheme also achieves a low paging cost and delay compared to other schemes proposed in the literature.

Location prediction is one of the key technologies of active mobility management in the next generation of mobile communication systems. Most of known location prediction models only take parts of predictive factors into account, which leads to a low prediction success ratio of these models. The motivation of this paper is to design a location prediction model considering multiple predictive factors to improve the prediction success ratio and improve the efficiency of the model. In this paper, a location prediction model based on Bayesian Network theory is proposed. The proposed model can effectively solve multi-factor location prediction. Firstly, the relative predictive factors are coded in the Bayesian Network node, and location prediction results can be calculated based on cell topology information integrated in the model structure. A factors distribution mechanism is designed to solve the problem when the nodes cannot obtain location prediction information directly. Subsequently, the method of calculating location prediction results for each cell is presented. The simulation results indicate that the proposed location prediction model is effective in improving accuracy of location prediction and the stability of the model is better than that in comparative schemes.

Network selection is an important issue in the next generation of heterogeneous wireless network and well studied for individual flows. However, researches of network selection for multi-flows at network side are seldom touched upon but also important from the global view of optimizing usage of network resources. In this paper, a weighted bipartite graph based network selection scheme for multi-flows is proposed, which adopt secondary exponential smoothing method to perform network resource prediction and assign flows to networks based on Matching Degree(MD). The network selection scheme is modeled as weighted bipartite graph with objectives of maximizing Matching Degrees and access ratio as well as the constraint of guaranteeing no network overloaded. Weighted Bipartite Graph Algorithm (WBGA) is designed to achieve the goals. Simulation results show that WBGA demonstrates best performance compared to other schemes, with regard to average MD, access ratio and utilization ratio of network resources.

In MIMO systems, spatial multiplexing is a powerful technique for increasing channel capacity by transmitting multiple data streams in the same channel simultaneously. Moreover, additional performance can be extracted in the presence of channel state information (CSI) at the transmitter. However, channel estimation error usually exists in practical systems and leads to imperfect CSI. As a result, the system performance is degraded. Fortunately power allocation can mitigate the problem effectively. In this paper, the power allocation problem is investigated in the case of imperfect CSI with accurate system model. A greedy power allocation (GPA) algorithm with adaptive modulation scheme is proposed to maximize the system data rate while satisfying each data stream's bit error rate requirement. Simulation results show that GPA can reduce the effects of imperfect CSI and obtain better performance than other traditional algorithms, e.g. waterfilling and equal power allocation algorithms.

Various paging strategies have been proposed to improve the efficiency of paging management. However, most of the schemes ignore how to obtain location predictions or are too complex for real systems. In this paper, we propose a new adaptive paging scheme designed according to the mobility pattern and location probabilities. The Bayesian network is used as the location prediction model which describes a broad class of mobility patterns. The probability distribution of a mobile terminal's location is derived on the condition that the incoming calls form a Poisson process and the cell holding time has an exponential probability distribution. The paging strategy is implemented using a heuristic algorithm which can adaptively change according to the given mobility pattern and traffic parameters. The running time of the strategy is only ¿(N) where N describes the number of cells in a Tracking Area. Experimental results show that the adaptive paging scheme also achieves a low paging cost compared to other schemes proposed in the literature.