Abstract

The natural medicines have complicated therapeutic behaviours, which are very responsive to environmental stressors, but the conventional methods of analysis are poor in order to capture the dynamic aspects of interplay. The work describes a smart biomedical biosensor system and Artificial Intelligence (AI) and evolutionary optimization toward assessing the therapeutic efficacy of natural medicines in different environmental conditions. A biosensor based on nanomaterials was developed and experimented under controlled temperature and humidity, pH, and exposure to pollutants to collect data on biochemical responses in high-resolution form. Gradient Boosted Regression Trees (XG Boost) was utilised to predict nonlinear associations among the bio-sensor responses, environmental, and therapeutic efficacy markers. The model had an excellent predictive (R 2 = 0.96) and was able to determine the determinants that have a significant effect on efficacy. To improve the sensor robustness, we employed multi-objective evolutionary optimization algorithm (NSGA-II), during which refinement was made on the geometry of the biosensor, the coating thickness, and biosensor operating conditions. The designs that were optimised showed significant gains in terms of sensitivity, response time, and stressor stability. The multi-biosensing-AI-optimization system presents a very precise, resilient, and scalable method towards analysing natural remedies. The work presents a new direction of smart biosensing solutions to personalised medicine, natural product authentication, and real time health observation of the environment.