Research On Optimization Scheme Of Multi-user Spectrum Allocation Based On Deep Reinforcement Learning In Wireless Body Area Networks

The advancement of science and technology and the improvement of living standards have promoted the improvement of the average life expectancy of human beings,but at the same time,the problem of population aging has gradually become prominent.According to research,my country's elderly population will reach 400 million in 2035,accounting for nearly 30%.The incidence of chronic diseases in the elderly has also increased,which can even be life-threatening without early detection and intervention.If real-time health monitoring of the human body can be carried out,and early diagnosis and treatment can be carried out in a timely manner,the probability of these diseases developing in the direction of deterioration will be reduced.The emergence and development of wireless body area network technology can be applied to such medical scenarios.Wireless Body Area Networks(WBAN)is a wireless network technology based on radio frequency,which belongs to a kind of wireless sensor network.A WBAN consists of a set of energy-efficient,miniaturized sensor nodes with wireless communication capabilities placed inside or on the body.By collecting important biological information of the human body,such as blood pressure,heart rate,blood oxygen,body temperature,etc.,it can be transmitted to the terminal device through Wi Fi,Bluetooth,etc.,and then to the cloud server for 24 hours without affecting human health and normal life.dynamic real-time monitoring.However,due to the limited spectrum resources in the body area network,and as the number of users accessing the network continues to increase,the user density in a certain area is too high,which will inevitably lead to network congestion,resulting in serious co-channel interference,which in turn affects the network.performance is negatively affected.The existing interference suppression technologies mainly include the processing of interference from outside the network and the coordination of interference within the network.Compared with the former,the latter requires lower hardware performance and complexity,and with the development of artificial intelligence technology,the It also has a good effect when applied to interference coordination in the network.Based on the above,aiming at the optimization problem of multi-user spectrum resource allocation in the body area network,this paper proposes a resource allocation algorithm based on deep reinforcement learning.The main work is as follows:(1)The communication model of the system is determined.It is assumed that each wireless body area network user has a central HUB and several sensor nodes.According to the function and importance of the nodes,each wireless The body area network user is transformed into a communication model in which a key node transmits to the HUB.In this network environment,a gateway is set to exist,which can control whether to access the channel by sending special instructions to the wireless body area network user.The gateway belongs to the second layer of the communication hierarchy in the wireless body area network,which has a certain hardware complexity and performance and can be used to run the proposed machine learning algorithm.(2)Application of reinforcement learning algorithm: the communication between sensor nodes and HUB,that is,channel access,becomes the action transfer function of reinforcement learning algorithm Q-learning,and energy efficiency is defined as the reward in reinforcement learning.The agent interacts with the network environment by,in the case of unpredictable real-time dynamic environment of the network,continuous interactive trial,trial and error,learning to obtain the best channel access strategy.(3)Application of deep reinforcement learning: On the basis of Q-learning,a deep learning network is introduced to realize deep reinforcement learning(Deep Reinforcement Learning),which combines the decision-making ability of reinforcement learning with the perception ability of deep learning.Perceptual decision making in the network environment.When the complexity of the environment is high,deep reinforcement learning solves the problem of state space explosion in traditional reinforcement learning.The simulation results show that,using the deep network instead of the list to fit the Q value,the performance of the network is improved,and the energy efficiency is further improved.The resource allocation algorithm based on deep reinforcement learning proposed in this paper is executed by the gateway,which controls the access channel of wireless body area network users,and obtains a nearly optimal resource allocation strategy by dynamically sensing the external environment.The inter-network interference suppression has achieved good results.