Research On Access Control Of Wireless Body Area Network With Sleep-Based Parallel Reservation Service

In recent years,wireless body area network(WBAN)has attracted a lot of attention from scholars for its efficiency and convenience,and it has been applied to the medical field.However,due to the limited energy of the nodes,critical data needs to be transmitted in the face of the diversity of medical applications,energy efficiency,low latency and quality of service(Qo S)have become one of the fundamental requirements of WBAN,so one of the challenges in designing WBAN is how to ensure good service quality while ensuring energy efficiency.In the WBAN communication,the access control strategy at the media access control(MAC)layer controls the access method and resource allocation of the nodes.And polling service can provide service guarantees as it avoids the conflict of competing channels between nodes.Based on this,this paper considers energy consumption and service performance,based on the polling service approach,the access control strategy for WBAN is studied,and the main work is as follows:(1)For node idle energy consumption problem,a parallel polling control strategy with a sleep mechanism(PPCS-S)is proposed in this paper.By adopting a sleep mechanism to reduce the idle waiting energy consumption of nodes,at the same time,use parallel threshold service method to improve the efficiency of data transmission,thereby enhances the energy efficiency of the whole system and extends the life cycle of the system.The paper also uses embedded Markov chains and probabilistic generating functions,established the mathematical model of PPCS-S,accurately derive performance characteristics such as the mean cyclic period and joint queue length at the polling epoch,etc.Further,the closed-form expressions for the quantitative relationship between sleep time and network service parameters are derived,and provide a theoretical approach for assessing the energy efficiency of a system.The simulation experimental results show that the theoretical values of the proposed model matched with the experimental results,compared to existing polling service strategies,service performance has been significantly improved,and energy consumption has also been reduced.(2)Due to the complexity of the mathematical model of the polling service system,the difficulty of its derivation and the difficulty of resolving performance metrics.When faced with different service strategies,the analysis of performance all require re-modeling of the problem.This paper combines reinforcement learning methods,an autonomous scheduling service access strategy based on Deep Q Network(DQN)technology is designed(DQN-P).In this scheme,the system service process continuously learns the relationship between network service parameters and system feature quantities.Thus,learning to obtain the network service parameters under the balance of system performance and energy consumption,to achieve dynamic adjustment of service strategy.The experimental results show that the DQN-P can control the service strategy of each site by learning to adjust in a timely manner,which is more flexible in taking into account the service system delay and service fairness.