体域网的信道分配协议及姿态识别研究

发布时间：2019-02-23 22:31

【摘要】：体域网由附着在人体体表、体内或周边环境的若干传感器节点组成,能够实时采集人体各项生理指标数据,并进行分析计算和传输。在医疗、体育赛场等场景中,每个使用者及周边的传感器节点构成一个体域网,多个使用者则构成一个更大规模的体域网。为使体域网内各节点间通信有效进行,应研究体域网的信道分配协议,主要涉及通信效率、通信公平性、功率控制等问题。此外,动作识别技术在医疗和体育领域也有广泛应用,微机电系统的发展提高了采用体域网进行动作识别的精度,降低了对环境的要求。以体域网为研究对象,针对节点信道分配协议的效率问题、公平性问题、功耗问题进行深入研究,同时采用微机电系统设计了基于体域网的姿态识别方法,主要工作如下：针对多信道环境中节点通信的信道分配效率问题,将博弈论引入信道分配算法中,提出了基于完全且完美信息动态博弈的信道分配算法。假定各节点均为理性并且贪婪,在节点选择信道时,采用逆向归纳方法得出全部其他节点的选择策略,以此为基础计算得到当前节点自身的优化选择策略,最终实现网络中信道的高效动态分配,提高了网络吞吐率,降低了丢包率和传输时延。针对节点获取信道资源的公平性问题,设计了基于信道占用率的自适应公平性算法。根据节点对通信历史及当前信道占用率与理想信道占用率之间的关系,将通信情况分为四种类别,再结合当前网络负载情况,动态设置竞争窗口值。该算法在改善吞吐率指标的同时,显著提高了节点信道分配公平性,避免了饥饿节点出现,在各种负载条件下表现良好。针对体域网节点发送信息的功耗问题,提出一种基于功率控制的媒体访问控制(MAC, Media Access Control)层信道分配算法,将不同功率值分配到不同信道,通信节点对根据距离选择数据包发送功率,进而选择该功率对应的信道。该方法有效降低了冲突域范围,减少了非对称连接造成的通信冲突,提高了网络总吞吐率。同时,该方法也降低了网络能耗,减小了节点发热,避免使人体受到灼伤。针对体域网中的姿态识别问题,设计了姿态识别节点,采用自适应扩展卡尔曼滤波方法进行姿态识别,利用节点的加速度计和电子罗盘数据对陀螺仪输出角速度进行补偿,对状态变量进行迭代计算,减少了系统误差影响。实验表明该方法提高了姿态识别的有效性。
[Abstract]:Body area network is composed of several sensor nodes attached to human body surface, body or surrounding environment. It can collect human body physiological index data in real time, analyze, calculate and transmit. In the medical and sports scene, each user and the surrounding sensor nodes form an individual domain network, while multiple users form a larger body area network. In order to make the communication between nodes in the body area network effective, we should study the channel assignment protocol of the body area network, which mainly involves communication efficiency, communication fairness, power control and so on. In addition, motion recognition technology is also widely used in medical and sports fields. With the development of MEMS, the accuracy of motion recognition using body area network is improved, and the requirement of environment is reduced. Taking the body area network as the research object, the efficiency, fairness and power consumption of the node channel assignment protocol are studied deeply. At the same time, the attitude recognition method based on the body area network is designed by using the micro electromechanical system (MEMS). The main work is as follows: aiming at the channel allocation efficiency of node communication in multi-channel environment, this paper introduces game theory into channel allocation algorithm, and proposes a channel allocation algorithm based on complete and perfect information dynamic game. Assuming that each node is rational and greedy, when the node selects the channel, the selection strategy of all other nodes is obtained by using the reverse induction method, and the optimal selection strategy of the current node itself is calculated based on this method. Finally, the efficient and dynamic channel allocation in the network is realized, which improves the throughput of the network and reduces the packet loss rate and transmission delay. An adaptive fairness algorithm based on channel occupancy is designed to solve the fairness problem. According to the history of node pair communication and the relationship between the current channel occupancy rate and the ideal channel occupancy rate, the communication situation is divided into four categories. Combining with the current network load, the contention window value is dynamically set. The algorithm not only improves the throughput index, but also significantly improves the fairness of the node channel allocation, avoids the emergence of hungry nodes, and performs well under various load conditions. In order to solve the problem of power consumption, a power control based media access control (MAC, Media Access Control) channel allocation algorithm is proposed, which allocates different power values to different channels. The communication node pair selects the transmission power of the packet according to the distance and then selects the channel corresponding to the power. This method can effectively reduce the range of conflict domain, reduce the communication conflicts caused by asymmetric connections, and improve the total throughput of the network. At the same time, this method also reduces the network energy consumption, reduces the node heat, and avoids the human body being burned. Aiming at the problem of attitude recognition in the body area network, the attitude recognition node is designed. The adaptive extended Kalman filter is used to recognize the attitude, and the accelerometer of the node and the electronic compass data are used to compensate the output angular velocity of the gyroscope. The state variables are calculated iteratively to reduce the influence of system error. Experiments show that this method improves the effectiveness of attitude recognition.
【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP391.44;TN929.5

【相似文献】