Smart Anomaly Detection Using Data-Driven Techniques in IoT Edge: A Survey

The Internet of Things (IoT) technology has begun to proliferate in recent years, which simultaneously increases the number of attacks. Owing to the massive volume and multi-dimensional data in IoT, anomaly detection leads to low prediction accuracy and a high false alarm rate. Further, there is a deficit of real-world test datasets for anomaly detection. This work aims to generate a novel real-time anomaly detection dataset and proposes an efficient anomaly detection model using an Improved Grey Wolf Optimization (IGWO)-enabled Long Short-Term Memory (LSTM) network in IoT edge scenarios. Dataset generation is carried out using a testbed setup containing Raspberry Pi 4 and sensors connected by a lightweight Message Queuing Telemetry Transport (MQTT) protocol. An autoencoder is used for feature reduction as it can investigate the input characteristics without sacrificing vital information. The LSTM classifier parameters, such as learning rate, optimizer, and batch size, are tuned precisely using IGWO techniques. The experimental results disclose that the proposed model achieves an accuracy of 99.11% for the testbed dataset, which is better than recent models. To confirm the generalizability of our model, the CICIDS 2017, DS2OS, and MQTTset standard datasets are applied explicitly. The developed model outcomes are statistically verified using the Wilcoxon signed rank test.INDEX TERMS Anomaly detection, autoencoder, improved grey wolf optimization, Internet of Things security, LSTM networks, MQTT, Wilcoxon signed rank test.

Section: ) Frame Work For Anomaly Detectionmentioning

confidence: 99%

DL-ADS: Improved Grey Wolf Optimization Enabled AE-LSTM Technique for Efficient Network Anomaly Detection in Internet of Thing Edge Computing

Manokaran,

Vairavel

2024

“…Anomaly detection(AD) or outlier detection refers to the techniques for identifying patterns or observations in data that do not conform to an expected behavior [131]. This techniques are used in variety of IoT smart applications [131] such as, smart city, smart monitoring and smart power management. Mostly of this applications used sensors as an input device.…”

Section: Tinyml and Anomaly Detectionmentioning

confidence: 99%

“…Mostly of this applications used sensors as an input device. These devices generated a huge volume of data that are transmitted to the cloud servers for analysis, decision making and storage [131].…”

Section: Tinyml and Anomaly Detectionmentioning

confidence: 99%

See 1 more Smart Citation

A Comprehensive Survey on TinyML

Abadade,

Temouden,

Bamoumen

et al. 2023

Recent spectacular progress in computational technologies has led to an unprecedented boom in the field of Artificial Intelligence (AI). AI is now used in a plethora of research areas and has demonstrated its capability to bring new approaches and solutions to various research problems. However, the extensive computation required to train AI algorithms comes with a cost. Driven by the need to reduce the energy consumption, the carbon footprint and the cost of computers running machine learning algorithms, TinyML is nowadays considered as a promising AI alternative focusing on technologies and applications for extremely low-profile devices. This paper presents the results of a literature survey of all TinyML applications and related research efforts. Our survey builds a taxonomy of TinyML techniques that have been used so far to bring new solutions to various domains, such as healthcare, smart farming, environment, and anomaly detection. Finally, this survey highlights the remaining challenges and points out possible future research directions. We anticipate that this survey will motivate further discussions on the various fields of applications of TinyML and the synergy of resource-constrained devices and edge intelligence.

“…In previous research, individual ML algorithms have been examined, but their false alarm rates and detection rates are not more effective. Many ML algorithms have been considered for building IDS, such as decision tree (DT) [11], dimensionality reduction algorithms [12], random forest (RF) [13], swarm intelligence techniques [14], support vector machine (SVM) [15], K-nearest neighbor (KNN) [16], logistic regression (LR) [16], and naive Bayes (NB) [17]. However, designing a robust anomaly detection model using a single ML algorithm is a challenging endeavor.…”

Section: Introductionmentioning

confidence: 99%

IGWO-SoE: Improved Grey Wolf Optimization Based Stack of Ensemble Learning Algorithm for Anomaly Detection in Internet of Things Edge Computing

Manokaran,

Vairavel

2023

Self Cite

With the tremendous growth and popularization of the Internet of Things (IoT), the number of attacks targeting such devices has also increased. Therefore, enhancing the anomaly detection model to maximize detection accuracy and mitigate cyber-attacks in time-critical IoT edge scenarios is essential. Furthermore, there is a lack of vivid, precise, cross-layered, and diverse datasets in IoT for evaluating these anomaly detection models. This paper aims to develop an improved anomaly detection model based on an optimized stacked ensemble learning algorithm at edge computing. Initially, a novel synthetic dataset with multiple cross-layer attacks is generated using the Cooja simulator to train our proposed model. In addition, by introducing an improved grey wolf optimization (IGWO) approach, the parameters of ensemble learning algorithms, such as the number of trees, learning rate, and sample rate, are tuned precisely, and the stacking ensemble concept is applied to the optimized ensemble learning algorithms to enhance their prediction capabilities. The experimental results demonstrate that the developed model produces a detection accuracy of 99.44% for our proposed Cooja simulated dataset, which is higher than the contemporary methods. The generalizability of the proposed model is expressed explicitly using four different datasets: NSL KDD, UNSW NB 15, MQTTset, and CICIDS 2017. Finally, we assess the befitting of the proposed model using a chi-square statistical significance test, thereby providing an enriched contribution to the recent works in anomaly detection.INDEX TERMS Anomaly detection, cooja simulator, edge computing, ensemble learning, improved grey wolf optimization, Internet of things, statistical significance test.