Synthesis of IoT Sensor Telemetry Data for Smart Home Edge-IDS Evaluation

Smart homes comprise of gadgets like refrigerators, air conditioners, energy meters etc, which send telemetry data to the cloud for analysis, decision making and control. Smart home networks are prone to attacks like denial of service, injection attack etc., which need to be detected by the Intrusion Detection Systems (IDS). The challenge in the development of Machine Learning (ML) based IDS is the scarcity of actual data for ML model generation and evaluation. In this paper, an approach of IDS based on the time difference between samples is proposed. Also, how the impact of the attacks can be synthesized, is described. An XGBoost based classifier is evaluated using this synthetic data. The quality of this synthetic data has been computed in terms of Training on Synthetic data and Testing on Real Data (TSTR) and Prediction Capability (PC). This synthetic data can be generated with multiple levels of Attack Impact Factor (AIF), where the level is determined by how difficult it is to classify the data accurately.