Secret Elliptic Curve-Based Bidirectional Gated Unit Assisted Residual Network for Enabling Secure IoT Data Transmission and Classification Using Blockchain

In recent days, a wide range of Internet of Things (IOT) related applications are employed for automated services. Various issues such as security, reliability and fault tolerance has restricted the use of IoT services in real time environments. Also, the transmitted data are prone to various types of attacks. Therefore, there is a need for secure data transmission and different types of attack has to be classified. Thus, a novel encryption technique is proposed in this study for affording data security in IoT system and a novel deep learning technique is proposed for the effective categorization of attacks in IoT data. Initially, the input data is hashed using Amended Merkle Tree approach (AMerT). Then, the data are encrypted using Secret Elliptic curve cryptography (SEllC). Next, the encrypted data are assembled in blockchain framework which is an Interplanetary File System (IPFS). Finally, decryption is carried out to access the given input data and the attacks are classified by proposing a novel deep learning technique called Attention Bidirectional Gated unit assisted Residual network (Att-BGR). Here, in order to improve the classification accuracy attention is used along with Bidirectional Gated unit assisted Residual network. Here, the Experimentation is carried out via Python and the efficiency of proposed work is determined by evaluating varied metrics. The comparison results exhibits that the proposed technique obtains better classification accuracy having an accuracy of 98.38%. Also, the encryption approach utilized has a minimum encryption time of 114s which is lower than the existing techniques.