A Study On Energy-efficient Telemonitoring System For Human Activity Based On WBANs

In recent years,the remote monitoring system based on the motion mode of low-power body-area networks has received great attention in the applieation of telemedicine,which can contribute to preventing tumble of the elderly,rehabilitation and treament of cerebrovascular chronic diseases,etc.The mode of low-power body-area networks and high-performance long-range motion pattern recognition is a hot issue in the related research fields.The main research work of this article are as follows:Considering the heterogeneity of collection nodes in body area network,based on superframes,this paper proposes a multi-factor weighted time slot allocation method.Under the master-slave star network architecture,this protocol adopts a time division multiplexing access control strategy based on superframes and proposes a dynamic time slot allocation mechanism.The method firstly sets multi-factor weights to perform initial time slot allocation,and initially determines the numbers and sequences of time slots.After many iterations of the four-factor weight,time complexity of the algorithm can be effectively reduced,and simultaneously,the nodes can be collected in the process of Super frame transmission.The time slot utilization rate is increased by parameter 1 in the case of maximum data delay in the unfilled time slots in the node set in this paper.Under the condition of guaranteeing delay and buffer limitation,taking full use of time slots,the number of information exchange is reduced to achieve the requirement of reducing energy consumption.A remote monitoring system based on low power consumption and wearable human motion mode is constructed.The system architecture consists of three parts:wearable domain network,Internet and remote data processing system.The basic idea is as follows:both of the compressed sensing algorithm and the low power consumption in above are integrated into the hardware design of domain network of wearable collection sensor node and the coordination control convergence,and a low energy agreement with multiple sensing movement data compression algorithm is proposed by combining the perception process data gathering strategy,to accurately acquire motion data remotely,identify the distal human movement mode and achieve remote human movement mode behavior monitoring.In this paper,A novel human motion pattern classification algorithm is proposed based on the second-order nearest neighbor sparse reconstruction which aims at accurately identifying the changes of human motion patterns and realizing the purpose of remote monitoring.The basic idea of the proposed algorithm is as follows:in view of the high dimensionality of multi-sensing motion data,this paper first investegate the near-neighbor sample sets of each type of neighbors for undetermined test samples,and construct the action classification model by means of fast neighboring sample set construction.The Lagrangian algorithm solves its cooperative representation coefficients,where the sparse coefficients are utilized to reconstruct the first sample and various types of reconstructed samples are obtained meanwhile,and then uses the neighboring principle to find a certain number of neighbor classes and neighbors for reconstructing samples.Additionally,the sample is reconstructed into a training sample set.Finally,a fast collaborative classification representation model is constructed based on the training sample set to solve the cooperating display coefficients and representation residuals of the pending test sample,determining the category to which the pending test sample belongs,and implementing action classification.The algorithm reduces the time complexity while improving the recognition rate,and provides a new method for identifying the human motion pattern.