Deep Q-Learning-Based Neural Network for Secure Data Transmission in Internet of Things (IoT) Healthcare Application

The Internet of Things (IoT) is revolutionizing the healthcare industry and expediting real-time monitoring through its vast connection between clinicians, patients, clinical and nursing staff, and medical equipment. Both possibilities and challenges for information sharing and gathering arise from the network’s scale and diversity. Ensuring patients’ safety and privacy requires safeguarding the medical equipment they use, with a focus on patient data such as health status. To analyze healthcare and treat patients on time, specialists provide consultation on sensitive patient data. Numerous uses of biometric and cryptographic systems exist, such as anomaly detection and authentication for medical systems and deep learning (DL) techniques for system security and authentication. The most important thing to remember when implementing a deep learning-based security system is that security and efficiency must be balanced because the network’s sensors are energy-constrained devices. Consequently, this study developed a unique framework, the deep Q-learning-based neural network, for secure data transmission in the IoT healthcare sector with shorter encryption and decryption times to protect data transmission from outside threats. Network traffic is reduced by processing patient data in this manner. This method also reduces expenses and mistakes in communication. When compared to other traditional approaches, the proposed model performed better, like recurrent neural network (RNN), Elman neural network (ENN), and deep neural network (DNN). The suggested deep Q-learn-based neural network (DQNN) approach specifically obtained 32.12 ms encryption time, 43.58 ms decryption time, 93.62% sensitivity, 91.47% specificity, and 94.75% accuracy.