​ Abstract: This work presents an Edge Artificial Biosensing Architecture that integrates electrochemical sensing with artificial intelligence for real-time physiological monitoring. The proposed system has integrated a glucose oxidase-based electrochemical biosensor with the analog signal conditioning, feature extraction, and a quantized CNN deployed on the Raspberry Pi 4B. The framework executes localized signal processing and inference for low latency and improved data privacy. Experimental evaluation on 5000 biosensor signal samples achieved a classification accuracy of 96.2 %, a low inference latency of 0.12 seconds, and a power consumption of 1.6 W. Finally, the cloud-assisted retraining mechanism resulted in a further enhancement of model performance by 1.8 %, enabling adaptability to environmental drifts and long-term operation. The results establish that integrating AI with edge computing greatly enhances the biosensor's efficiency, scalability, and autonomy, while opening up new horizons for next-generation intelligent biosensing and personalized healthcare in the IoMT domain.

Keywords: Artificial intelligence, Edge computing, Electrochemical biosensor, Convolutional Neural Network (CNN), Glucose Oxidase (GOx), Signal conditioning, Internet of Medical Things (IoMT), Quantized deep learning, Cloud retraining, Intelligent healthcare.

_________________________________________________________________________________________