基于时间序列预测的体域网数据融合算法研究

发布时间：2018-06-28 05:14

本文选题：体域网 + 数据预测　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：无线通信技术的发展彻底改变了我们的日常生活,其技术应用涉及自动化控制、跟踪与监控。随着无线传感器网络(Wireless Sensor Network,WSN)技术的发展改进,出现了体域网(Wireless Body Area Network,WBAN),通过将低成本的、能量受限的、微小的异构传感器节点穿戴甚至植入人体,形成一个特殊的型无线传感器网络,采集生理信息,通过无线通信到中央基站,实现了对人体各个指标的实时监测。WBAN传感器设备提供实时反馈,对人体不会造成任何不适,从而为用户提供更大灵活性和流动性,可以代替复杂有线医疗设备,能够连续监测人体的重要生理信号和娱乐信号。与所有其他电子系统相同,WBAN也需要适当的电源保证其正常运作。目前,WBAN大都采用电池供电系统。但相比较其他WSN系统,由于传感器网络设计及应用的特殊性,WBAN电池难以及时更换,所以应用各种节能技术,为WBAN提供一个非常长的网络寿命,是研究的重点领域之一。一般来讲,WBAN节能主要采用使用更加节能的电子器件、减少能源消耗的设备数量以及采用更加有效的算法,使得对于一个给定的电池容量,提高在系统的操作过程中的能量使用效率。WBAN能量的消耗可以分为三个部分:感知、无线通信和数据处理,其中无线通信是最耗电的。体域网所收集的信息一般都具有较高的时间冗余和空间冗余。通过数据融合技术,减少WBAN中的无线通信量可以达到节省能量的目的。根据以上分析本文进行了以下工作:(1)本文提出了一种基于小波分析和最小二乘支持向量机的时间序列数据预测技术的数据融合算法。通过在感知节点和汇聚节点处分别建立一个相同的轻量级数据预测模型,当预测结果在误差范围之内则不进行数据的传输,从而降低了网内冗余数据的传输,减少传输能耗。(2)本文研究了预测模型的参数优化的方法,并根据理论分析提出了一种改进的粒子群算法,该算法通过对基本粒子群算法中的参数进行改进以及结合遗传算法思想对算法迭代过程进行改进克服了原算法的一些不足。(3)本文基于对人体基本生理数据进行健康监测出发设计了一种WBAN系统,对系统的应用场景及功能进行了分析,设计了其架构,并对传感器节点以及网关节点进行了软硬件设计。最后将本文提出的算法在该实验平台上进行了验证分析。
[Abstract]:The development of wireless communication technology has completely changed our daily life, its technical application involves automation control, tracking and monitoring. With the development and improvement of Wireless Sensor Network (WSN) technology, Wireless body Area Network (WBAN) has emerged, by wearing or even implanting low-cost, energy-constrained, tiny heterogeneous sensor nodes into the human body. A special type of wireless sensor network is formed, which collects physiological information and wirelessly communicates to the central base station, which provides real-time feedback to the real-time monitoring of every index of the human body. WBAN sensor equipment does not cause any discomfort to the human body. Thus, it can provide more flexibility and mobility for users, can replace complex wired medical equipment, and can continuously monitor important physiological and entertainment signals of human body. WBAN, like all other electronic systems, also requires proper power supply to ensure its proper operation. At present, most of WBAN use battery power supply system. However, compared with other WSN systems, because of the particularity of sensor network design and application, WBAN battery is difficult to be replaced in time, so the application of various energy-saving technologies to provide a very long network life for WBAN is one of the key areas of research. In general, WBAN uses more energy efficient electronic devices, reduces the number of energy consuming devices, and uses more efficient algorithms, so that for a given battery capacity, The energy consumption of WBAN can be divided into three parts: perception, wireless communication and data processing, among which wireless communication is the most power consuming. The information collected by body area network generally has high time redundancy and space redundancy. By using data fusion technology, the energy saving can be achieved by reducing the wireless traffic in WBAN. Based on the above analysis, the following works have been done: (1) this paper proposes a data fusion algorithm based on wavelet analysis and least squares support vector machine (LS-SVM) for time series data prediction. By establishing the same lightweight data prediction model at the perceptual node and the convergence node respectively, when the prediction results are within the error range, the data transmission is not carried out, thus reducing the transmission of redundant data in the network. (2) the method of parameter optimization of prediction model is studied in this paper, and an improved particle swarm optimization algorithm is proposed according to theoretical analysis. The algorithm overcomes some shortcomings of the original algorithm by improving the parameters in the basic particle swarm optimization algorithm and the iterative process of the algorithm with the idea of genetic algorithm. (3) based on the basic physiological data of human body, this paper is based on the basic physiological data. A WBAN system is designed for Kang Monitoring. The application scene and function of the system are analyzed, the structure of the system is designed, and the hardware and software of the sensor node and the gateway node are designed. Finally, the proposed algorithm is verified and analyzed on the platform.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5

【参考文献】