Deep Reinforcement Learning Optimization for the Body Centric Sensor Network Design

The healthcare sector is currently undergoing a digital transformation, leveraging advancements in technology across various domains such as biomedical sensors, cloud computing, big data, and Internet of Things (IoT). This transformation involves the gradual integration of these digital components into the healthcare industry. The body-centric sensor network (BCSN) is strategically deployed within the patient’s body to continuously gather and monitor various biological functions, including respiration, blood pressure, temperature, heart rate, etc. In this paper, we study the BCSN design problem for future healthcare systems by considering key parameters such as physiological measurement requirements, patient body characteristics, and sensor features. Then, we propose a deep reinforcement learning algorithm, named DDPG-BBS, to solve the studied problem. It balances the implementation of biomedical sensor resources with the collection of healthcare data. We evaluate the performance of this design method through simulations involving different healthcare cases. The analysis explores the trade-offs between healthcare performance and resource investment based on the evaluation results. Simulation results show that the DDPG-BBS algorithm is superior to the baseline algorithm and it can efficiently support next-generation healthcare systems.