Many studies employ deep learning methods to process remote sensing data for large-scale yield estimation, often focusing on extracting growth features from the input images while overlooking regional variations in yield label distribution. An issue is that the resolution and overhead view of these images limit their ability to fully capture the growth environment, potentially constraining model performance when relying solely on image features. To address this, we propose LM-net, a novel predictive model with two key components. First, a linear temporal feature (LTF) block models historical yield as a time series, capturing yield distributions and regional differences in growth conditions. Second, a multi-scale attention (MSA) block processes remote sensing data to extract correlations at various temporal resolutions throughout the crop growth season. Embeddings from these blocks are combined and fed into a multi-layer perceptron decoder for yield estimation. Extensive experiments show that our model outperforms state-of-the-art methods on multiple metrics. Moreover, incorporating the LTF block into existing models significantly improves yield estimation accuracy, demonstrating its applicability and effectiveness.

The three-layer cellular network integrating mobile cloud computing (MCC) and mobile edge computing (MEC) provides an effective example for emerging services with intensive computing and low latency requirements. Computing intensive and delay sensitive tasks can be processed by local, edge nodes or cloud services. In the paper, we consider a three-tier collaborative computing network composed of mobile devices, edge nodes and cloud servers. In this network, we jointly optimize the offloading decision, the computation, communication and caching (3C) resource to minimize task delay. However, the formulated problem is a large-scale mixed integer nonlinear optimization problem with the increasing number of base stations and devices, which is difficult to solve. To deal with the problem, we propose a parallel processing scheme based on alternating direction method of multipliers (ADMM) and relaxation method. The proposed scheme first reconstructs the original problem through relaxation and auxiliary variables, then decomposes the large-scale problem into several sub problems and improves the task processing efficiency through parallel processing. Simulation results show that the proposed scheme can achieve near optimal performance under low complexity, and can reduce the task delay about 26% compared with other schemes.

Agricultural multi-robot task allocation is an important direction of intelligent agricultural robots. In this paper, to ensure the shortest total movement distance of robots and balance the workload, we transform the task allocation of agricultural multi-robot into a multi objective multiple traveling salesman problem (MO-MTSP). A processing framework is also designed to adjust the allocation plan in real-time by monitoring the status of robots and tasks. In this processing framework, an improved Non-dominated genetic algorithm (INSGA-II) is utilized to achieve better performance of task allocation. The improvement includes: 1) An objective function is designed to ensure a balanced task workload distribution for each robot considering the actual conditions of the farmland; 2) A neighborhood search algorithm is used to accelerate the convergence of the algorithm by adjusting the initial population plan; 3) An adaptive adjustment mechanism for crossover and mutation probabilities is proposed to enhance the flexibility of the task allocation plan searching. Finally, the proposed algorithm is evaluated on both public datasets and actual agricultural plot datasets. Experimental results show that the improved algorithm can achieve better allocation results, demonstrating the practical applicability of the algorithm in task allocation of multiple robots employed in agriculture scenarios.

Since low delay is required in most healthcare user cases, data computing are performed by mobile edge computing (MEC) deployed in Radio Access Networks (RANs). Therefore, MEC computing resource must be allocated among terminals properly. In existing researches, MEC computing resource is allocated dynamically or statically. However, dynamic schemes suffer from high complexity and static schemes will lead to poor service quality or low resource utilization. To achieve tradeoff between complexity and efficiency, this paper designs a semi-dynamic scheme for MEC computing resource allocation, called SD-CRA. As its main idea, MEC computing resource is firstly assigned to access point (AP) which is then shared by its users dynamically. And computing resource allocated to AP is updated periodically. However, updating period in SDCRA lasts longer than that in dynamic schemes since AP’s computing tasks, which equals sum of its users’ computing tasks, varies much slower than user’s. To design SD-CRA, we consider wireless transmitting adequately since computing tasks need be transmitted to MEC wirelessly. With this consideration, one minimal problem of MEC computing resource is formulated. To solve the formulated problem, research scenarios are first divided into two categories based on wireless channel capacity. In ideal scenario of unlimited wireless channel capacity, algebraic expression of MEC computing resource allocation is provided with its applicable condition. And in scenario of limited wireless channel capacity, two optimal problems are then designed considering the tradeoff between minimal MEC computing resource and maximal wireless throughput, which are solved by proposed algorithms. Performances of complexity and efficiency of proposed algorithms are expressed by simulation results.

The mobile edge computing (MEC)-enabled cellular network provides a promising paradigm for emerging services with intensive computation and low delay requirement. However, when multiple computation-intensive and delay-sensitive services are required concurrently, how to minimize the delay and monetary cost on mobile devices (MDs) under the limited computation resource and communication resource remains a challenging issue. In order to solve the issue, we formulate computation offloading, communication resource and computation resource allocation as an optimization problem in MEC-enabled cellular network. The problem is non-convex. Hence, we formulate it as a multi-objective computation offloading and resource allocation (MCORA) game and prove the existence of Nash equilibrium (NE). To obtain the NE, we design a MCORA algorithm taking the uplink, downlink spectrum resource, computation resource and offloading strategy into consideration. Simulation results show the MCORA algorithm can minimize system cost of all MDs. Moreover, in comparison to the existing algorithms, the proposed algorithm can achieve better performance in delay and system cost.

This paper introduces a network coding-based protection scheme against single- and multiple-link failures. The proposed strategy ensures that in a connection, each node receives two copies of the same data unit: one copy on the working circuit and a second copy that can be extracted from linear combinations of data units transmitted on a shared protection path. This guarantees instantaneous recovery of data units upon the failure of a working circuit. The strategy can be implemented at an overlay layer, which makes its deployment simple and scalable. While the proposed strategy is similar in spirit to the work of Kamal in 2007 2010, there are significant differences. In particular, it provides protection against multiple-link failures. The new scheme is simpler, less expensive, and does not require the synchronization required by the original scheme. The sharing of the protection circuit by a number of connections is the key to the reduction of the cost of protection. This paper also conducts a comparison of the cost of the proposed scheme to the 1+1 and shared backup path protection (SBPP) strategies and establishes the benefits of our strategy.

The p-cycle and its Failure Independent Path Protection (FIPP) extension are known to be efficient and agile protection strategies. The p-cycle is preconfigured such that if there is a failure, only the switches at two end nodes need to be reconfigured. In this paper, we extend the p-cycle by allowing cycles to have attached links, called Parasitic Protection Links (PPL), in order to protect paths whose source and destination nodes are not only located on the cycle but also connected by the PPL to the cycle. A p-cycle with PPL is named p2-cycle. We address the unicast service protection problems against single-link failures by using p2-cycle in mesh networks and the problem is formulated as an Integer Linear Program (ILP). The numerical results show that the p2-cycle scheme provides better capacity efficiency than the FIPP p-cycle scheme in all the traffic scenarios considered and consumes 2.7%-14.8% extra total cost over the optimum, provided by Shared Backup Path Protection (SBPP) approach. Moreover, we study the failure recovery performance by comparing it to FIPP and SBPP in terms of average number of reconfigurations (NOR). The results achieved by the p2-cycle is less than that of SBPP in every scenario considered and the number reduces from 2.48 to 2.31 comparing to 2, achieved by FIPP, as the traffic demands increases in NSFNET. In conclusion, the p2-cycle provides greater overall performance over existent p-cycle-based and path-based protection schemes, especially when the number of traffic demands is large.

In this paper we address the problem of traffic grooming in WDM rings with all-to-all and its generalization to many-to-many service by using network coding. We consider minimizing the number of Line Terminating Equipment (LTE) on two types of unidirectional rings, namely, single-hub and un-hubbed rings, as our objective. In single-hub rings, we investigate the minimum cost provisioning of uniform all-to-all traffic in two cases: where network coding is used to linearly combine data, and where it is not used and data is transmitted without coding. We generalize the service mode to many-to-many and evaluate the cost of provisioning. In un-hubbed ring, we propose a multi-hub approach to obtain the minimum cost provisioning in the case of all-to-all and many-to-many traffic. From our numerical results, we find that under many to many traffic, network coding can reduce the network cost by 10-20% in single-hub rings and 1-5% in un-hubbed rings.

As the demand for broadband wireless access services grows rapidly, much attention has been paid to integrated communication and broadcast networks (ICBN). ICBN can provide high-rate, on-demand, reliable data and multimedia services while in fast moving. In this paper, we introduce an ICBN system in which DMB-T broadcast network works as the forward link and wireless access networks are used for the return channel respectively. We first differentiate four types of services, including digital TV, video on demand, data download and Internet access. Then the system features are given. Next, we design the system architecture and the three subsystems, including the DMB-T broadcast network, the return channel and the terminal. With the introduction of subnetwork data unit (SNDU), the protocol to transmit data on DMB-T broadcast network is presented. We also stack the functions of the system into different layers in terms of four types of services, and introduce each of them, including signaling messages exchange and ARQ scheme, etc. Then the testbed is briefly introduced. Finally we draw the conclusion.