Anomaly Detection in Internet of Things using Clustering Technique

Abstract

Widely scattered sensing devices gather data from target regions, which are then transmitted to a processing centre for aggregation and analysis. The quality of services is typically determined by the accuracy and integrity of the data. However, the abnormal data will be gathered because of the unfavourable environment or device flaws. The network is extremely susceptible to network anomalies, especially black hole and misdirection assaults. However, it's not always easy to spot active attacks, particularly for remote sensing applications. The ability to detect anomalies accurately is essential for guaranteeing service quality. In the suggested hybrid anomaly detection method for misdirection and black hole assaults, the K-medoid tailored clustering algorithm is utilised. By specifying network parameters and collecting threshold values to find anomalies, a synthetic dataset was produced. Testing was carried out using the network simulator (NS-2).