Multi_key Homomorphic Crypto and Multiscale Transformer Learning Secure Data Aggregation in Wsn for Smart Agriculture

Two paramount factors pivotal for agricultural Internet of Things (IoT) systems are security and data aggregation, making certain effective management of resources, informed decision-making, and safety protection of sensitive data as far as Wireless Sensor Networks (WSNs) are concerned. On one hand, data aggregation encompasses integrating data from numerous sensors, and on the other hand, security ensures the confidentiality and integrity of data. These two paramount factors are convoluted in smart agriculture applications and must be focused on to establish trust between farmers. Initiating strong data privacy mechanisms and protocols and instructing farmers about the significance of data security are indispensable steps. Secure algorithms for sharing data and appropriately aggregating the data can take the edge off these issues. To address the above-said research gap in this work, a method called Multi-key Partial Homomorphic Cryptography and Multi-scale Data Aggregation-based Transformer Learning (MPHC-MDATL) is proposed. The MPHC-MDATL method is split into two sections, namely, cryptography to ensure data security, and then securely aggregating data in WSN for smart agriculture. Initially, Multi-key Partial Homomorphic Encryption-based Cryptography for the security model is applied to the IoT Agriculture 2024 dataset. Next, aggregation of data is performed by using Multi-scale Data Aggregation-based Transformer Learning without revealing the underlying sensitive data. Finally, to get the original digits of data packets from the ciphertext, Multi-key Partial Homomorphic Decryption-based Cryptography is applied. After selecting a secure mechanism with improved data confidentiality and data integrity, the data aggregation process is carried out with minimal delay, energy consumption, and a higher packet delivery ratio. We performed experimental assessments with different factors, namely, data confidentiality, data integrity, end-to-end delay, energy consumption, and packet delivery ratio. The research outcome of the proposed MPHC-MDATL method conducted on 20% of improved energy consumption, 9% of packet delivery ratio, 14% and 10% of data confidentiality and integrity, with 8% of minimum energy consumption than the state-of-the-art.