Quick and Accurate False Data Detection in Mobile Crowd Sensing

Li, Xiaocan; Xie, Kun; Wang, Xin; Xie, Gaogang; Xie, Dong; Li, Zhenyu; Wen, Junhai; Diao, Zulong

doi:10.1109/tnet.2020.2982685

Cited by 22 publications

(5 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, the work in [9] tackles data poisoning attacks on MCS platforms and the authors present the impact of data poisoning attacks on several machine learning classifiers when poisoning data are injected by the adversaries that are equipped with capabilities to build and run AI models. The study in [10] proposes a Lightweight Low Rank and False Matrix Separation (LightLRFMS) scheme to accelerate the computation time to detect false sensing data at significantly higher accuracy levels in comparison to the other separation / decomposition techniques.…”

Section: Related Work and Motivationmentioning

confidence: 99%

Generative Adversarial Network-Driven Detection of Adversarial Tasks in Mobile Crowdsensing

Chen,

Kantarci

2022

Preprint

View full text Add to dashboard Cite

Mobile Crowdsensing systems are vulnerable to various attacks as they build on non-dedicated and ubiquitous properties. Machine learning (ML)-based approaches are widely investigated to build attack detection systems and ensure MCS systems security. However, adversaries that aim to clog the sensing front-end and MCS back-end leverage intelligent techniques, which are challenging for MCS platform and service providers to develop appropriate detection frameworks against these attacks. Generative Adversarial Networks (GANs) have been applied to generate synthetic samples, that are extremely similar to the real ones, deceiving classifiers such that the synthetic samples are indistinguishable from the originals. Previous works suggest that GAN-based attacks exhibit more crucial devastation than empirically designed attack samples, and result in low detection rate at the MCS platform. With this in mind, this paper aims to detect intelligently designed illegitimate sensing service requests by integrating a GAN-based model. To this end, we propose a twolevel cascading classifier that combines the GAN discriminator with a binary classifier to prevent adversarial fake tasks. Through simulations, we compare our results to a single-level binary classifier, and the numeric results show that proposed approach raises Adversarial Attack Detection Rate (AADR), from 0% to 97.5% by KNN/NB, from 45.9% to 100% by Decision Tree. Meanwhile, with two-levels classifiers, Original Attack Detection Rate (OADR) improves for the three binary classifiers, with comparison, such as NB from 26.1% to 61.5%.

show abstract

Section: Related Work and Motivationmentioning

confidence: 99%

Generative Adversarial Network-Driven Detection of Adversarial Tasks in Mobile Crowdsensing

Chen,

Kantarci

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Most recent studies, however, mainly focus on challenges in the context of sensing task allocation, sparse sensing, privacy, and data integrity [6]. But only a few works try to address the problem of false data detection [7], [8]. For instance, there are data verification techniques such as Deco [8] which uses spatio-temporal techniques to reconstruct missing values.…”

Section: Related Workmentioning

confidence: 99%

“…Methods surrounding matrix separation have also been successfully used for separating false data. Light Weight Low Rank and Sparse Matrix Separation (LightLRFMS) [7] is a matrix separation technique but it was only validated on a specific MCS context for environmental monitoring that does not fully consider issues pertaining to concept-drift, which FSD strives to address based on the use of LSTMs. Furthermore, various deep learning techniques have previously been applied in MSC including LSTM models towards predicting traffic flow [10] or user mobility [11].…”

Section: Related Workmentioning

confidence: 99%

A Perfect Match: Deep Learning Towards Enhanced Data Trustworthiness in Crowd-Sensing Systems

Afzal-Houshmand

Homayoun

Giannetsos

2021

2021 IEEE International Mediterranean Conference on Communications and Networking (MeditCom)

View full text Add to dashboard Cite

The advent of IoT edge devices has enabled the collection of rich datasets, as part of Mobile Crowd Sensing (MCS), which has emerged as a key enabler for a wide gamut of safetycritical applications ranging from traffic control, environmental monitoring to assistive healthcare. Despite the clear advantages that such unprecedented quantity of data brings forth, it is also subject to inherent data trustworthiness challenges due to factors such as malevolent input and faulty sensors. Compounding this issue, there has been a plethora of proposed solutions, based on the use of traditional machine learning algorithms, towards assessing and sifting faulty data without any assumption on the trustworthiness of their source. However, there are still a number of open issues: how to cope with the presence of strong, colluding adversaries while at the same time efficiently managing this high influx of incoming user data. In this work, we meet these challenges by proposing the hybrid use of Deep Learning schemes (i.e., LSTMs) and conventional Machine Learning classifiers (i.e. One-Class Classifiers) for detecting and filtering out false data points. We provide a prototype implementation coupled with a detailed performance evaluation under various (attack) scenarios, employing both real and synthetic datasets. Our results showcase how the proposed solution outperforms various existing resilient aggregation and outlier detection schemes.

show abstract

“…Paper [40] used graph theory tree structure to construct an undirected spanning tree, and then broadcasted the route through the central node in the cluster, and searched first to get a minimum route spanning tree, and then used the algorithm of ant colony to traverse the minimum spanning tree, and finally got the optimal route in the cluster. Paper [41] divided the monitoring area into several sub-areas, looking for the highest energy node in the sub-area as the cluster head node, and the routing link formed by the cluster head nodes in all sub-areas. The above operation was repeated after N cycles, then a new routing link can be found, and finally the maximum optimized route for data transmission can be completed.…”

Section: B Optimizes Clusteringmentioning

confidence: 99%

CoC-SCS: Cooperative-Optimization Coverage Algorithm Based on Sensor Cloud Systems in Intelligent Computing

Sun

2020

IEEE Access

View full text Add to dashboard Cite

the coverage of traditional Wireless Sensor Networks (WSNs) is limited by node energy and data redundancy, which forces WSNs to be interrupted abnormally. Therefore, Cooperative-optimization Coverage Algorithm based on Sensor Cloud Systems in Intelligent Computing (CoC-SCS) is proposed. First, the algorithm determines the location information of the Focus Target Nodes (FTNs), and uses the Genetic Algorithm (GA) to give the node path planning. Second, the mutation parameters and the controllable threshold parameters are used in order to control the event domain nodes. Optimizing clustering makes the clustering of the nodes to be more uniform, so that it can improve the search ability of the global target nodes and even reduce the nodes energy consumption. Third, the adaptation function covers the continuity of the covered target positions and the monitoring range of the nodes. Optimization is also performed to achieve the goals of increasing network coverage and extending the network lifetime. Finally, simulation results show that the proposed CoC-SCS algorithm compared with other three algorithms in this paper has improved 11.27% and 16.35% on average in terms of network coverage, network lifetime, etc., and 13.95% in terms of network energy overhead, thereby we can further verify that CoC-SCS algorithm has strong stability and effectiveness.

show abstract

Quick and Accurate False Data Detection in Mobile Crowd Sensing

Cited by 22 publications

References 33 publications

Generative Adversarial Network-Driven Detection of Adversarial Tasks in Mobile Crowdsensing

Generative Adversarial Network-Driven Detection of Adversarial Tasks in Mobile Crowdsensing

A Perfect Match: Deep Learning Towards Enhanced Data Trustworthiness in Crowd-Sensing Systems

CoC-SCS: Cooperative-Optimization Coverage Algorithm Based on Sensor Cloud Systems in Intelligent Computing

Contact Info

Product

Resources

About