Super-resolving remotely acquired hyperspectral images, often with low resolution (LR), is a critical signal processing technique, as it greatly affects the subsequent material classification and identification tasks. An economical approach in the remote sensing area is to fuse the spatial details extracted from the high-resolution (HR) counterpart multispectral image into the LR hyperspectral image, thereby inferring the desired HR hyperspectral image. Convex analysis has been shown to be an effective tool for the fusion mission, but it often relies on sophisticated regularization schemes to tackle this challenging inverse problem. In the existing literature, the deep plug-and-play strategy was proposed for fast implementation of those sophisticated regularizers, but just approximately without convergence guarantees. Thus, we introduce deep learning (in an alternative approach) to tailor a simple convex regularizer for efficient super-resolution. Remarkably, though typical deep fusion methods can tackle non-linear effect presented in real hyperspectral data, they often rely on big data and sophisticated network structures, which are often time-consuming and resource-intensive. Instead, our deep regularizer just needs a small-data-driven simple network architecture that implies better stability and tractability; we achieve so by reconsidering the role of deep learning as simply to guide the convex algorithm to search the fusion solution, rather than directly serving as the final solution. The proposed convex deep image fusion (CODE-IF) algorithm, with all the closed-form algorithmic expressions derived, achieves state-of-the-art hyperspectral super-resolution performance.

Considering that hyperspectral image (HSI) is often of lower spatial resolution when compared to multispectral image (MSI), an economical approach for obtaining a high-spatial-resolution (HSR) HSI is to fuse the acquired HSI and MSI, thereby greatly facilitating the subsequent material identification and classification in satellite remote sensing. As satellite-acquired HSI and MSI are often misaligned, the proposed deep neural network does not require the input HSI/MSI to be spatially co-registered, making the challenging fusion network design even more difficult. In this study, we propose a streamlined and efficient convex model integrated into the sub-network, which obviates the need for complex network structures in learning spatial-spectral relationships, effectively guiding the quasi-residual learning task in our alignment-free fusion network. The convex sub-network is a low-rank model that leverages the convex geometric structure implicitly embedded in the hyperspectral signature space. To address the misalignment between HSI and MSI effectively, we introduce a novel Shifted Window Attention Module (SWAM) that exploits the neighboring correlation in the feature domain, significantly enhancing the performance and stability of the fusion task. Capitalizing on the redundancy among spectrums, we employ grouped convolution to decrease the computational complexity without causing additional performance degradation. The proposed Quasi-residual Convex Deep Network (QRCODE) demonstrates state-of-the-art performance in alignment-free HSI/MSI fusion tasks.

In 5G, network functions can be scaled out/in dynamically to adjust the capacity for network slices. The scale-out/-in procedure, namely autoscaling, enhances performance by scaling out instances and reduces operational costs by scaling in instances. However, the autoscaling problems in 5G networks are different from those in traditional cloud computing. The 5G network functions must be considered the simultaneous deployment of multiple instances; moreover, the deployment of 5G network functions is more frequent than that of traditional cloud computing. Both the number and timing of deployment will substantially affect the cost-effectiveness of the system. In this paper, we first identify the autoscaling issues specifically based on the 3GPP standards. We develop a low-complexity analytical queuing model to formulate the problem and quantify a set of performance metrics with closed-form solutions. The proposed analytical model and closed-form solutions are cross-validated by extensive simulations. The analytical model offers design insights and theoretical guidelines, helping us study the effectiveness of reservations. We proposed a dynamic block-setup reservation algorithm (DBRA) to find the optimal reserved number and threshold value of network slices. Therefore, mobile operators can balance the system's cost-effectiveness without large-scaled testing and real deployment, saving cost on time and money.

In 5G, multi-access edge computing enables the applications to be offloaded to near-end edge servers for faster response. According to the 3GPP standards, users in 5G are separated into many types, e.g., vehicles, AR/VR, IoT devices, etc. Specifically, the high-priority traffic can preempt edge resources to guarantee the service quality. However, even if a traffic is transmitted with low priority, its latency requirement in 5G is much lower than that in 4G. Too strict latency requirement and priority-based service make resource configuration difficult on the edge side. Therefore, we propose the edge-cloud offloading mechanism, in which each edge server can offload tasks to back-end cloud server to ensure service quality of both high- and low-priority traffic. In this paper, we establish a priority-based queuing system to model the edge-cloud offloading behaviors. Based on the formulation of our system model, we propose Knapsack Potential Game (KPG) to derive an optimal offloading ratio for each edge server to balance the cost-effectiveness of the overall system. We demonstrate that KPG has low computational complexity and outperforms two baseline algorithms. The results indicate that KPG’s performance is optimal and provides a theoretical guideline to operators while designing their edge-cloud offloading strategies without large-scale implementation.

Cellular network control procedures (e.g., mobility, idle-active transition to conserve energy) directly influence data plane behavior, impacting user-experienced delay. Recognizing this control-data plane interdependence, L25GC re-architects the 5G Core (5GC) network, and its processing, to reduce latency of control plane operations and their impact on the data plane. Exploiting shared memory, L25GC eliminates message serialization and HTTP processing overheads, while being 3GPP-standards compliant. We improve data plane processing by factoring the functions to avoid control-data plane interference, and using scalable, flow-level packet classifiers for forwarding-rule lookups. Utilizing buffers at the 5GC, L25GC implements paging, and an intelligent handover scheme avoiding 3GPP's hairpin routing, and data loss caused by limited buffering at 5G base stations, reduces delay and unnecessary message processing. L25GC's integrated failure resiliency transparently recovers from failures of 5GC software network functions and hardware much faster than 3GPP's reattach recovery procedure. L25GC is built based on free5GC, an open-source kernel-based 5GC implementation. L25GC reduces event completion time by ~50% for several control plane events and improves data packet latency (due to improved control plane communication) by ~2×, during paging and handover events, compared to free5GC. L25GC's design is general, although current implementation supports a limited number of user sessions.

In 5G mobile networks, users can request services from different network slices in the core network. However, in 3GPP Release 16 (R16), Mutually Exclusive Access to Network Slices (MEANS) was introduced, in which some services may not be accessed simultaneously by the same user. Because the users do not know which slices are mutually incompatible, and the core network has no specific way to deal with it, the mechanisms defined in the standards will not only cause signaling overhead but waste additional time. In this paper, we propose Adaptively Prioritized Subset Algorithm (APSA) as a solution to MEANS, which follows the principles discussed in the 3GPP documents. We conducted an extensive simulation to evaluate the proposed APSA. The results show that our proposed APSA not only uses fewer signaling messages to access the network slices but also outperforms other algorithms compared in this paper.

In 5G, the requirement of transmission latency is stricter than that in 4G. To enhance transmission efficiency, a user plane function (UPF) with a specific packet processing mechanism is necessary. However, UPF must communicate with the session management function (SMF), which will send the packet processing rules to UPF. Those rules will substantially occupy UPF storage. Moreover, customizing a UPF needs to reconstruct N3, N4, N6, and N9 interfaces, which takes much time for developers. To this end, we propose the user plane function development framework (UPFDF), which modularizes the functions in the UPF, supporting customization to connect different types of packet processing mechanisms. With UPFDF, we address the UPF capacity problem and improve the flexibility of the system.

Next generation intelligent transportation systems aim at many cooperative perception and cooperative driving functions that need significant computational resources. Offloading such tasks to some mobile edge computing solutions is considered part of the solution, which is currently investigated in the scope of 5G networks. In the automotive context, such edge systems could be road-side units (RSU), which, however, can easily be overloaded at peak times. Vehicular micro-cloud approaches have been proposed to overcome such problems by sharing computational resources of nearby cars. In this study, we propose an offloading system architecture to enable such offloading such vehicular micro-cloud interconnected by a 5G core network. We model the system as a queueing model to derive closed-form solutions for selected performance metrics. Based on these insights, we propose the Double-Check Offloading Algorithm (DCOA) to obtain the best offloading ratio to the vehicular micro-cloud. Our simulation results show the proposed DCOA has better system performances compared with four other offloading schemes.

One of the main goal of 5G networks is to provide ultra-reliable low latency service to users. When users keep accessing the system, the traffic in 5G core network (5GC) may be congested. Thus, we propose a load balance algorithm to select the best traffic data routing path based on the traffic loading in the 5GC. The proposed algorithm is implemented in a 5G testbed called free5GC. The experimental results show that our proposed algorithm outperforms the traditional round-robin load balance algorithm in many performance metrics.

Inspired by the self-exploring learning approach, this paper proposes an object-learning system in which a robot interacts with objects to obtain their features and construct object concepts. The system consists of three kinds of features: interaction features, visual features, and intrinsic features. When the robot interacts with an object, it observes the changes in the object to obtain its interaction features. At the same time, the robot learns the visual features of the object. The intrinsic features are the properties of the object. Models of the relationships among the three kinds of features are constructed through an Artificial Bee Colony based Random Forest algorithm and a Convolutional Neural Network. The established models help the robot to predict the properties of new objects and to make decisions. Two experiments are constructed in this paper: the service-providing task and the stacking task. In the former, the robot decides on an appropriate object, using the object concept models, to accomplish an appointed task. In the second experiment, the robot uses a Gated Recurrent Neural Network to learn the stacking sequence of various objects. All the experimental results demonstrate that the robot can build object concept models by interacting with objects, and can utilize these models to accomplish various tasks.