Fault diagnosis in clustering WSN based on neighbor cooperation

Zhang, Ying; Xian-feng, Yuan

doi:10.1109/ccdc.2016.7531274

Cited by 9 publications

(6 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A data fault is a sensor data report that is inconsistent with the actual behavior of the phenomenon of interest [7]. Several fault detection schemes have been proposed to address this problem [8] [9]. Nevertheless, this type of faults is beyond the scope of this investigation.…”

Section: Network Fault Monitoringmentioning

confidence: 99%

“…Dividing nodes into clusters facilitates distributed computing such as routing and data aggregation. Several distributed fault diagnosis techniques previously proposed often rely on some forms of clustering mechanisms [22] [23] [24] [25] [26]. Zhang and Yuan [22] proposed a fault diagnosis algorithm that divides the whole network into several clusters so that fault management can be distributed into each cluster area.…”

Section: Region-based Monitoring Techniquementioning

confidence: 99%

See 1 more Smart Citation

Distributed Region-Based Monitoring in Low-Power Listening Wireless Sensor Networks

Pattamasiriwat

Jaikaeo

2022

ECTI-CIT Transactions

View full text Add to dashboard Cite

Advancement in IoT technology and the concept of Information-Centric Networking lead to less importance of node individuality since several nodes can work interchangeably. Multiple sensor nodes can be grouped into a region and monitored as one instance to guarantee su°cient coverage over the region. Therefore, a single node fault often does not need to be reported unless it is the last node in the region. In addition, applications focusing on detecting rare events rarely require nodes to transmit and often rely on a low-power listening MAC protocols, where nodes spend most of their time sleeping but require signi˝cantly more work during transmission. In such situations it is desirable to avoid periodic status reports transmitted to the central monitor station as usually found in a centralized monitoring scheme. A distributed region-based monitoring scheme, or DRMON, is then proposed to facilitate this circumstance. This approach designates a representative to each region so that it can be used as an indicator of the region's status with a mechanism to re-elect a new representative until all nodes in the respective region are dead, implying region inactiveness. We evaluate the suitability of DRMON over various scenarios in two aspects: centralized vs. distributed monitoring schemes and individual-based vs. region-based monitoring schemes, along with existing work in the literature. Simulation results indicate that region-based schemes outperform the individual schemes in terms of power consumption and scalability when the number of regions is low. The distributed schemes also yield better e°-ciency in terms of message overhead. Compared against the other schemes, DRMON's overall power consumption is reduced by 4%-10%, with 66%- 88% reduction in packet transmissions, while maintaining fault detection precision and recall of greater than 90% and the detection delay within an acceptable range. This outcome suggests that in the case where existence of individual node is out of concern, distributed region-based fault monitoring scheme could be employed to reduce energy usage and lower message overhead while retaining acceptable detection accuracy.

show abstract

Section: Network Fault Monitoringmentioning

confidence: 99%

Section: Region-based Monitoring Techniquementioning

confidence: 99%

Distributed Region-Based Monitoring in Low-Power Listening Wireless Sensor Networks

Pattamasiriwat

Jaikaeo

2022

ECTI-CIT Transactions

View full text Add to dashboard Cite

show abstract

“…A data fault is a sensor data report that is inconsistent with the actual behavior of the phenomenon of interest [6]. Several fault detection schemes have been proposed to address this problem [7] [8]. Nevertheless, this type of faults is beyond the scope of this investigation.…”

Section: Network Fault Monitoringmentioning

confidence: 99%

Distributed Region-Based Monitoring in Low-Power Listening Wireless Sensor Networks

Pattamasiriwat

Jaikaeo

2020

Preprint

View full text Add to dashboard Cite

Advancement in IoT technology and the concept of Information-Centric Networking lead to less importance of node individuality since several nodes can work interchangeably. Multiple sensor nodes can be grouped into a region and monitored as one instance to guarantee sufficient coverage over the region. Therefore, a single node fault often does not need to be reported unless it is the last node in the region. In addition, there are occasions where a central monitor station cannot rely on continuous data delivery from nodes or regions to decide whether they are still alive, such as situations when nodes are deployed to detect rare events. Moreover, low-power listening MAC protocols, which significantly help reducing power consumption while nodes are mostly idle, put a lot more work on the transmission process. In such situations it is desirable to minimize status reports to the central monitor station. A distributed region-based monitoring scheme, or DRMON, is then proposed to facilitate this circumstance. This approach designates a representative to each region so that it can be used as an indicator of the region's status with a mechanism to re-elect a new representative until all nodes in the respective region are dead, implying region inactiveness. We evaluate the suitability of DRMON over various scenarios in two aspects: centralized vs. distributed monitoring schemes and individual-based vs. region-based monitoring schemes. Simulation results indicate that region-based schemes outperform the individual schemes in terms of power consumption and scalability when the number of regions is low. The distributed schemes also yield better efficiency in terms of message overhead and load distribution. In addition, detection accuracy of all schemes is not significantly different and fault detection delay is guaranteed. This outcome suggests that in the case where existence of individual node is out of concern, distributed region-based fault monitoring scheme could be employed to reduce energy usage and lower message overhead while retaining the detection accuracy.

show abstract

“…The diffusion estimation algorithm uses the robust function such as Huber's cost function and error saturation using non linearity for making network fault tolerant. Zhang et al [22] discusses the concept of fault diagnosis in WSNs using a clustering technique which is based on distributed energy-efficient clustering algorithm as well as neighbor cooperation strategy. The algorithm splits whole network into various clusters, thus fault management is dispersed to each cluster area, and results in the improved neighbor collaborative diagnosis strategy within the cluster adopted.…”

Section: Literature Reviewmentioning

confidence: 99%

“…The average of the all the recommendations collected after a particular Δt time is the Rechp and then the relative trust metric deviation is as in Equation (22).…”

Section: Estimation Of Network Layer Trustmentioning

confidence: 99%

Cross Layer Security Based Intrusion Detection System (CLS-IDS) For Securing Wireless Sensor Network (WSN)

Nandala¹,

Sumalatha²

2021

DSA

View full text Add to dashboard Cite

The Wireless Sensor Network (WSN) consists of large number of tiny sensor nodes that are deployed randomly in the harsh or unattended environment. Generally, the sensor nodes are prone to unexpected failures and faults as they are resource constrained. Therefore, to maximize the network lifetime and to meet the QoS requirements by analyzing faults, cross layer design is considered to be best optimization technique to improve overall network performance. Previous research work developed a cross layer security based intrusion detection system (CLS-IDS) framework that has improved network performance, minimal energy consumption and to prolong network life, meeting QoS requirements. However, this CLS-IDS method cannot detect the faults happening at various places at the same time. In addition, when the cross layer design method focuses only on the inter-layer interaction and the change of the node, it may not be able to adjust to dynamic change in network conditions. This cross layer design is predominantly vulnerable to various types of attacks at various protocol stack layers. This proves that it is essential to provide effective security based intrusion detection system and fault monitoring techniques. The proposed work focuses on the Cross Layer Security based Intrusion Detection System (CLS-IDS) framework, which is considered to be a vital constraint in escalating the network performance. The security framework considers both the trust value and the fault value to identify the malicious nodes based on the pre-defined trust threshold value. The fault monitoring system uses the Improved Support Vector Machine (ISVM) classifier to isolate and remove fault nodes. The experimental results show that proposed CLS-IDS perform better when compared to existing cross layer protocol designs.

show abstract

Fault diagnosis in clustering WSN based on neighbor cooperation

Cited by 9 publications

References 13 publications

Distributed Region-Based Monitoring in Low-Power Listening Wireless Sensor Networks

Distributed Region-Based Monitoring in Low-Power Listening Wireless Sensor Networks

Distributed Region-Based Monitoring in Low-Power Listening Wireless Sensor Networks

Cross Layer Security Based Intrusion Detection System (CLS-IDS) For Securing Wireless Sensor Network (WSN)

Contact Info

Product

Resources

About