Secured Heterogeneous Data Fusion Using Adversarial Autoencoders in Internet of Military Things

Internet of Military Things (IoMT) is the intervention of IoT technology in the Military applications to gain more knowledge on the unmanned areas, the battlefield, and for warfare purposes. As such, IoT, IoMT also uses sensors, communication networks and protocols, and userspecific application processes. However, the IoMT requires more security and integrity in the data collected from the sensor network, as a huge amount of data is required to predict the scenario on which a decision has to be made. On the other hand, as the volume of data increases, it reduces the QoS of the whole system. Heterogeneous data fusion plays a vital role in representing multi-modal data into a singular format that has to be replicable. Data from various sources need to be securely fused so that the used information is highly reliable. Therefore, Adversarial Autoencoder is proposed for heterogeneous data fusion in the IoMT application to reduce the data dimensionality with good integrity and security. The proposed work outperforms the conventional method in terms of evaluation parameters and efficiently detects the threats using the anomaly detection using the IR Small dataset and EMI dataset from Kaggle.