Research On The Inter-network Interference Management Mechanism Based On Coexistence State Prediction In WBANs

Wireless body area network(WBAN)is one of the main research fields in the Internet of Things(IoT).As an emerging technology,it is closely related to human body,and has huge demand and significant meaning for more and more widespread applications,such as sports,entertainments and medical services.Unlike other traditional networks,the WBAN is a human-centered wearable wireless network with sensor nodes deployed in/on/around a human body,and these nodes are used to monitor every kind of physiological information.Because of human mobility,the coexistence problem may occur when multiple WBAN meet with each other.It will influence the timeliness and reliability of data transmission,degrade the communication performance,and even threaten people's life.Therefore,in order to solve the coexistence problem in WBANs,this paper focuses on the inter-network management mechanism in the multi-WBAN environment,to guarantee the quality of service of the networks in complex situations.To ensure the coexistence problem could be detected and handled in time,this paper presents a machine learning-based method for coexistence state prediction in multiple WBANs,which divides the coexistence states into four classes:None,Static,Semi-dynamic and Dynamic.Supervised learning and feature extraction are utilized in the method to generate a set of models with various configurations,and the best one with the highest accuracy for prediction is got.Meanwhile,the experimental data used for training and testing models are acquired through CC2530 2.4GHz low power transceivers.Simulation results show that compared with another coexistence state prediction method,the proposed method has better performance in terms of accuracy and timeliness.What's more,this paper presents an inter-network interference management mechanism based on coexistence state prediction in multiple WBANs.Firstly the mechanism uses the prediction method to get the current coexistence state,then it chooses a suitable interference management methods for each different coexistence state,so that it can avoid or mitigate inter-network interference,reduce packet error rate(PER)and energy consumption,and maximize the utility of each network.To be more specific,it chooses channel hopping for the Dynamic and Semi-dynamic states,and a designed non-cooperative game-based distributed power control algorithm for the Static state,but it will not handle the None state.Comparing the proposed inter-network interference management mechanism with the other three mechanisms,it proves that the proposed mechanism is superior in terms of average PER,average power,system total utility,average payoff and average iterations.