Affected by hardware and wireless conditions in WSNs, raw sensory data usually have notable loss corruption. Existing studies mainly consider the interpolation of random missing absence There is also no strategy to handle successive data. To address these problems, this paper proposes a novel approach based on matrix completion (MC) recover corrupted By analyzing large set weather collected from 196 sensors Zhu Zhou, China, we verify that features low-rank, temporal stability, spatial...

Detecting anomalous traffic is a critical task for advanced Internet management. Many anomaly detection algorithms have been proposed recently. However, constrained by their matrix-based data model, existing often suffer from low accuracy in detection. To fully utilize the multi-dimensional information hidden data, this paper takes initiative to investigate potential and methodologies of performing tensor factorization more accurate More specifically, we model as three-way formulate problem...

Traffic anomaly detection is critical for advanced Internet management. Existing algorithms generally convert the high-dimensional data to a long vector, which compromises accuracy due loss of spatial information data. Moreover, they are designed based on separation normal and anomalous in time period, not only introduces high storage computation cost but also prevents timely anomalies. Online accurate traffic difficult support. To address challenge, this paper directly models monitoring...

The inference of traffic volume the whole network from partial measurements becomes increasingly critical for various engineering tasks, such as capacity planning and anomaly detection. Previous studies indicate that matrix completion is a possible solution this problem. However, 2-D cannot sufficiently capture spatial-temporal features data, these approaches fail to work when data missing ratio high. To fully exploit hidden structures paper models 3-way tensor formulates recovery problem...

Cooperative communication for wireless networks has gained a lot of recent interests due to its ability mitigate fading with exploration spatial diversity. The objective this paper is design an efficient algorithm minimize the total consumed power network while guaranteeing transmission reliability multiple active pairs through cooperative communications. This problem not been studied and much more challenging than relay assignment considered in literature work which simply targets reduce...

Simultaneous wireless information and power transfer (SWIPT) transmits powers nodes with the same radio frequency signal. It can prolong life time of energy-constrained nodes. Current works SWIPT focus on one-hop two-hop network. In order to verify performance in multi-hop network (MECWN) where energy harvested by receiver node be as an compensation for data forwarding, this paper concurrently considers routing selection MECWN. To reduce consumption, we first formulate allocation problem...

Matrix completion has emerged very recently and provides a new venue for low cost data gathering in Wireless Sensor Networks (WSNs). Existing schemes often assume that the matrix known fixed low-rank, which is unlikely to hold practical system environment monitoring. Environmental vary temporal spatial domains. By analyzing large set of weather collected from 196 sensors ZhuZhou, China, we reveal have features stability, relative rank stability. Taking advantage these features, propose an...

End-to-end network monitoring is essential to ensure transmission quality for Internet applications. However, in large-scale networks, full-mesh measurement of performance between all pairs infeasible. As a newly emerging sparsity representation technique, matrix completion allows the recovery low-rank using only small number random samples. Existing schemes often fix samples assuming rank known, while data features thus vary over time. In this paper, we propose exploit techniques derive...

The inference of traffic volume the whole network from partial measurements becomes increasingly critical for various engineering tasks, such as prediction, optimization, and anomaly detection. Previous studies indicate that matrix completion is a possible solution this problem. However, two-dimension cannot sufficiently capture spatial-temporal features data, these approaches fail to work when data missing ratio high. To fully exploit hidden structures paper models 3-way tensor formulates...

A major factor that prevents the large scale deployment of Mobile Crowd Sensing (MCS) is its sensing and communication cost. Given spatio-temporal correlation among environment monitoring data, matrix completion (MC) can be exploited to only monitor a small part locations time, infer remaining data. Rather than taking random measurements following basic MC theory, further reduce cost MCS while ensuring quality missing data inference, we propose an Active Sparse (AS-MCS) scheme which includes...

There are a lot of recent interests on cooperative communication (CC) in wireless networks. Despite the large capacity gain CC small networks with its capability mitigating fading taking advantage spatial diversity, can result severe interference and even degraded throughput. The aim this work is to concurrently exploit multi-radio multi-channel (MRMC) technique transmission combat co-channel improve performance multi-hop network. Our proposed solution considers routing, channel assignment,...

Cooperative communication (CC) for wireless networks has gained a lot of recent interests. It been shown that CC the potential to significantly increase capacity networks, with its ability mitigating fading by exploiting spatial diversity. However, most works on are limited single radio network. To demonstrate benefits in multiradio multihop network, this paper studies joint problem cooperative routing and relay assignment maximize minimum rate among set concurrent sessions. We first model...

Offloading computation intensive applications to mobile cloud is promising for overcoming the problems of limited computational resources and energy devices. However, without considering competition relationship users cloudlets in computing system, existing studies lack an incentive mechanism system achieve efficient application offloading resource provisioning. In this paper, we design MPTMG, a Multi-dimensional Pricing based on Two-sided Market Game. We propose three types prices:...

The inference of the network traffic matrix from partial measurement data becomes increasingly critical for various engineering tasks, such as capacity planning, load balancing, path setup, provisioning, anomaly detection, and failure recovery. recent study shows it is promising to more accurately interpolate missing with a 3-D tensor compared interpolation methods based on 2-D matrix. Despite potential, difficult form measurements taken at varying rate in practical network. To address...

Detecting anomalous traffic is a crucial task of managing networks. Many anomaly detection algorithms have been proposed recently. However, constrained by their matrix-based data model, existing often suffer from low accuracy. To fully utilize the multi-dimensional information hidden in data, this paper takes an initiative to investigate potential and methodologies performing tensor factorization for more accurate Internet detection. Only considering low-rank linearity features current...

The inference of the network traffic data from partial measurements becomes increasingly critical for various engineering tasks. By exploiting multi-dimensional structure, tensor completion is a promising technique more accurate missing inference. However, existing algorithms generally have strong assumption that global low-rank and try to find single model fit whole tensor. In practical system, subset may stronger correlation. this work, we propose novel localized (LTC) increase recovery...

Network trouble shooting, failure location, and anomaly detection rely heavily on network traffic measurement data. Due to the lack of infrastructure, high cost, unavoidable transmission loss, monitoring systems suffer from problem that data are incomplete. This article models as a tensor exploit its strong ability feature extraction recover missing Different traditional completion which relies factorization, we design novel Deep Adversarial Tensor Completion (DATC) scheme based Learning...

The inference of the network traffic matrix from partial measurement data becomes increasingly critical for various engineering tasks, such as capacity planning, load balancing, path setup, provisioning, anomaly detection, and failure recovery. recent study shows it is promising to more accurately interpolate missing with a three-dimensional tensor compared interpolation methods based on two-dimensional matrix. Despite potential, difficult form measurements taken at varying rate in practical...

Matrix completion has emerged very recently and provides a new venue for low cost data gathering in WSNs. Existing schemes often assume that the matrix known fixed low-rank, which is unlikely to hold practical monitoring system such as weather gathering. Weather varies temporal spatial domain with time. By analyzing large set of collected from 196 sensors ZhuZhou, China, we reveal have features stability, relative rank stability. Taking advantage these features, propose an on-line scheme...

The attacks, faults, and severe communication/system conditions in Mobile Crowd Sensing (MCS) make false data detection a critical problem. Observing the intrinsic low dimensionality of general monitoring sparsity data, can be performed based on separation normal anomalies. Although existing algorithm Direct Robust Matrix Factorization (DRMF) is proven to effective, requiring iteratively performing Singular Value Decomposition (SVD) for low-rank matrix approximation would result...

Monitoring the performance of a large network is very costly. Instead, subset paths or time intervals can be measured while inferring remaining data by leveraging their spatiotemporal correlations. The quality missing recovery highly relies on inference algorithms. Tensor completion has attracted some recent attentions with its capability exploiting multi-dimensional structure for more accurate inference. However, current tensor algorithms only model three-order interaction features through...