Enhanced Dynamic S-Box Design Based on Adaptive Heuristic Evolution and Multi-Stage Nonlinearity for Securing IoT-Based Remote Health Monitoring Systems

This paper presents an enhanced method for designing dynamic S-Boxes, integrating adaptive heuristic evolution with multi-stage nonlinearity enhancements, tailored to improve cryptographic security in IoT-based remote health monitoring systems. The unprecedented rise in patient numbers during the COVID-19 pandemic highlighted the need for continuous health monitoring, enabled by IoT devices. Given the sensitivity and volume of health data transmitted, ensuring robust data security has become essential. The proposed S-Box design introduces an adaptive mechanism that adjusts the permutation and substitution processes based on real-time cryptographic performance feedback, optimizing the S-Box structure dynamically. By incorporating multi-stage nonlinearity, the method increases cryptographic complexity, effectively countering linear and differential cryptanalysis. The hybrid approach combines modular operations, affine transformations, and enhanced multiplicative inversion, generating a broad spectrum of robust, key-dependent S-Boxes. The method is further enhanced by introducing lightweight block encryption to secure sensitive health data in resource-constrained IoT environments. This approach also extends cryptographic evaluations to include dynamic avalanche criteria and an enhanced bit independence metric, ensuring superior resilience against attacks. Experimental results demonstrate the effectiveness of the K-star classification method in predicting critical health conditions, achieving an accuracy of 95%, with lightweight encryption providing secure data transmission. Comparative studies validate the method’s superior performance over contemporary S-Box designs, emphasizing its potential for integration in high-security cryptographic frameworks, particularly for secure communication channels and image encryption in healthcare IoT systems.