Federated learning (FL) enables clients to collaboratively learn a shared task while keeping data privacy, which can be adopted at the edge of wireless networks improve intelligence. In this letter, we aim minimize training latency FL system for given loss by client scheduling. Instead assuming that prior information about channel state and local computing power is available, consider more practical scenario without knowing information. We first reformulate scheduling problem as multi-armed...

Federated learning (FL) is promising in enabling large-scale model training by massive devices without exposing their local datasets. However, due to limited wireless resources, traditional cloud-based FL system suffers from the bottleneck of communication overhead core network. Fueled this issue, we consider a hierarchical and formulate joint problem edge aggregation interval control resource allocation minimize weighted sum loss latency. To quantify performance, an upper bound average...

Federated learning (FL) framework enables user devices to collaboratively train a global model based on their local data sets without privacy leak. However, the training performance of FL is degraded when distributions different are incongruent. Fueled by this issue, we consider clustered (CFL) method where divided into several clusters according and trained simultaneously. Convergence analysis conducted, which shows that depends cosine similarity, device number per cluster, participation...

With the development of deep learning, fingerprints recognition based on neural networks is a widely used method in indoor localization. In this paper, we build long short-term memory (LSTM) recurrent neuron network to make regression between and locations order track moving target. Simulations are BLE5.0 environment use received signal strength indication (RSSI) as element fingerprints. Since preparation an inevitable time-consuming process testing phase LSTM, propose two methods improve...

Power Internet of Things (PIoT) is a promising solution to meet the increasing electricity demand modern cities, but real-time processing and analysis huge data collected by devices challengeable due limited computing capability long distance from cloud center. In this paper, we consider edge assisted PIoT where tasks can be either processed locally devices, or offloaded servers. Aiming maximize long-term system utility which defined as weighted sum reduction in latency energy consumption,...

Federated learning (FL) has become a promising solution to train shared model without exchanging local training samples. However, in the traditional cloud-based FL framework, clients suffer from limited energy budget and generate excessive communication overhead on backbone network. These drawbacks motivate us propose an energy-aware hierarchical federated framework which edge servers assist cloud server migrate models clients. Then joint computing power control client association problem is...

In this paper, we first employ non-orthogonal multiple access (NOMA) in a cognitive radio (CR) based mobile edge computing (MEC) network to reduce the system delay. particular, each secondary user has two computation tasks which can be offloaded MEC servers via NOMA, while transmissions of different users are orthogonal by exploiting spectrum bands. Then, investigate delay minimization problem jointly optimizing offloading policy, transmit power for offloading, capabilities local and...

The Industrial Internet of Things (IIoT) is emerging as a promising technology that can accelerate the application industrial intelligence to smart factories. Because sensitive nature user data, federated learning (FL) which performs distributed machine while preserving data privacy, leveraged meet accuracy and privacy requirements IIoT end devices/clients. However, unreliable communications in may result possible single-point failures typical single-server FL framework, thereby negatively...

Federated edge learning (FEEL) emerges as a privacy-preserving paradigm to effectively integrate computing for the implementation of deep learning-based vehicular applications. Nevertheless, incentive mechanism vehicles participating in varied tasks, has not been well explored yet. In this paper, software-defined network (SDN) technology is adopted training control among vehicles, and novel FEEL framework, namely SDN-assisted semi-decentralized (SSD-FEEL) investigated, where multiple servers...

Driven by its agile maneuverability and deployment, the unmanned aerial vehicle (UAV) becomes a potential enabler of terrestrial networks. In this paper, we consider downlink communications in UAV-assisted wireless communication network, where multi-antenna UAV assists ground base station (GBS) to forward signals multiple user equipments (UEs). The is associated with GBS through in-band backhaul, which shares spectrum resource access links between UEs UAV. optimization problem formulated...

There are many methods used for modeling of emergency response capability. Stochastic Petri net is a better choice because the strict mathematical definitions and analysis methodology. In order to dynamic evaluate capability, paper introduced stochastic process. Taking chemical accident as an example, analyzed its structure, obtained some important measures performance The research result shows that methodology simple reasonable capability evaluation, which can greatly help fire commanders...

Poyang lake of Jiangxi province, China, is the biggest wetland China. It one largest bases migrant birds especially in winter. The nature ecosystem region has relationship with birds' activities, and vice versa. Research important ways for knowing ecosystem. We attempt to set up a system surveillance estimation bird types amounts from video. background landscape wild field, mainly including lake, grass, sky. video camera handy one. Because such as Cygnuspsilas nature, they are likely settle...

Managing for local data among the increasingly popular consumer smart electronics in a secure manner while increasing user experience is still hard work. Federated learning (FL), can develop intelligent electronics-related applications protecting privacy of data. However, considering traditional cloud-based FL training process, with limited energy budget reduce efficiency. Hence, this paper, to total latency also meeting targeted minimum value loss function and long-term consumption...

A counter example is given to contradict an impressive result published recently about the stability of a parallel packet switch with bufferless input demultiplexors. While it was claimed in [Denis et al., June 2001] that number switching planes K should be at least 2[R/r] - 1 make PPS stable, we give stable = [R/r] + 1, where R and r rate external internal ports respectively. To justify validness example, introduce new mathematical model demultiplexors PPS, which would useful investigation...

Federated learning allows multiple clients to cooperatively train an efficient model without exposing clients' local data. However, this distributed training technique is susceptible Byzantine attack that obstruct the of global by altering or uploading erroneous gradients. To solve issue, in paper, we propose novel committee based Byzantine-tolerant federated algorithm (CBTFL), which guarantees resilience, convergence, and correctness training. A established CBTFL review uploaded gradients...

Emergency response and rescue process in chemical accidents is the representative of discrete event dynamic system. In order to evaluate emergency capability accidents, paper introduced Petri net for modeling analysis analyzed its structure, obtained some important measures performances process. The research result shows that methodology simple reasonable evaluation, which can greatly help fire commanders identify weaknesses optimize teams accidents.