基于信道响应的室内设备无关被动人体定位研究

发布时间：2018-09-06 12:59

【摘要】：随着物联网的飞速发展,无线定位成为当今新兴技术之一,其能够帮助智能系统实现基于位置的智能服务。相比于传统的主动定位,新型的无线设备无关被动定位能够无需用户携带任何相关电子设备而对用户的出现、位置甚至身份进行识别,该项技术能够被广泛应用于工厂、家庭场所的安全防护、设备保护、人员管理等领域。室内环境下传统的无线信号强度由于其自身的粗粒度性,在遭受室内多径效应的影响下无法准确地感知人体的存在,导致室内设备无关被动定位性能无法达到令人满意的程度。当今时代,WLAN技术得到了迅猛发展。其中,IEEE802.11 a/g/n协议中所采用的正交频繁复用技术为无线定位提供了载波水平的细粒度信道响应信息。信道响应包含了多载波水平上的信道状态信息,能够刻画室内多径传播特征,为室内细粒度高精度的设备无关被动定位发展提供了新的机遇。目前,基于信道响应的设备无关被动定位研究在国际上尚处于萌芽阶段,大量的基础问题尚且存在未得到解决,其中包括多载波信道状态信息用于室内设备无关被动定位领域的有效性、稳定性和高效性。本文课题探索利用细粒度多载波信道状态信息来取得先进的室内设备无关被动定位技术,推动我国无线室内设备无关被动定位技术的发展。本文从细粒度单通信链路设备无关被动人体定位的角度出发,主要针对以下几个方面开展研究:首先,为了降低设备无关被动人体检测现场勘测的开销花费,提高设备无关被动人体检测普适应用性,本文进行了大量的对比实验,提出利用振幅响应信息来量化无线链路对于人体移动敏感度,构建轻量级设备无关被动人体检测模型,实现自适应设备无关被动人体检测,降低现场勘测的开销。大量的科学实验结果表明自适应设备无关被动人体检测精度能够达到低于5%的误报率和漏报率,从而证明利用振幅响应信息感知人体移动的有效性和高效性。其次,针对振幅响应对于室内人体慢速移动感知能力较低的现象,提出了基于相位响应的免校验设备无关被动人体移动检测。在将网卡中原始的随机相位信息经过线性变换后,系统提取可用的稳定相位信息。本文研发了基于时域相位变异系数的轻量级实时无校验的设备无关被动人体检测模型,能够降低系统勘测开销,提升设备无关被动人体检测的普适性。大量的实验结果表明该检测模型能够取得低于5%的误报率和90%以上的检测准确度,同时也证明了相位响应信息在静态环境下的稳定性以及对于人体移动检测的有效性、高效性。再次,在利用振幅响应信息来实现对接收机无线感知能力的量化评估,从而帮助选择具有高感知度的接收机位置的基础上,提出基于贝叶斯分类原理的细粒度设备无关被动人体定位模型,提升室内无线设备无关被动定位的范围和精度,降低监测区域内盲点数量。多个场地的实验评估结果表明当接收机位于高无线感知度位置时,其定位精度相比于传统的定位技术能够提升21%的性能,这也证明信道响应信息应用于设备无关被动人体位置估计的有效性。最后,基于不同频率信道状态信息的差异性和相关性,提出了高精度的室内设备无关被动定位模型。本文工作充分利用信道响应特有的室内频率选择性衰减特性,研究两种新颖的单链路条件下的设备无关被动人体定位算法:加权贝叶斯定位和最大相似矩阵定位算法,通过利用有监督学习技术进一步提升室内设备无关被动人体定位精度。大量的实验评估表明该模型能够取得少于1米的平均定位误差,也反映了信道响应在设备无关被动人体定位的稳定性与高效性。
[Abstract]:With the rapid development of the Internet of Things, wireless location has become one of the emerging technologies, which can help intelligent systems to implement location-based intelligent services. Compared with traditional active location, the new wireless device-independent passive location can be used to locate users without any related electronic devices and to locate, even identify users. Recognition, the technology can be widely used in factories, home safety protection, equipment protection, personnel management and other fields. Because of its coarse-grained nature, the traditional wireless signal strength in indoor environment can not accurately perceive the presence of human body under the influence of indoor multipath effect, resulting in indoor equipment independent passive positioning. Nowadays, WLAN technology is developing rapidly. The OFDM technology used in IEEE802.11 a/g/n protocol provides fine-grained channel response information at carrier level for wireless location. Channel response includes channel state information at multi-carrier level and can describe indoor environment. Multipath propagation characteristics provide a new opportunity for the development of fine-grained and high-precision indoor device-independent passive positioning. Currently, the research on channel response-based device-independent passive positioning is still in its infancy in the world, and a large number of basic problems remain unsolved, including the use of multi-carrier channel state information for indoor equipment. Efficiency, stability and efficiency in the field of independent passive location. This paper explores the use of fine-grained multi-carrier channel state information to obtain advanced indoor device-independent passive location technology, and promotes the development of wireless indoor device-independent passive location technology in China. In order to reduce the cost of equipment-independent passive human body detection and improve the universal applicability of equipment-independent passive human body detection, a large number of comparative experiments are carried out in this paper, and amplitude response information is proposed to quantify the mobile of wireless link to human body. A lightweight device-independent passive human body detection model is constructed to realize adaptive device-independent passive human body detection and reduce the cost of field survey. Secondly, aiming at the low perception ability of amplitude response to indoor slow motion, a phase response-free device-independent passive human motion detection is proposed. After linear transformation of the original random phase information in the network card, the system extracts the available stable phase information. This paper develops a lightweight real-time device-independent passive human detection model based on time-domain phase variation coefficients, which can reduce the system survey overhead and improve the universality of device-independent passive human detection. Thirdly, the amplitude response information is used to quantify the radio perception ability of the receiver, which helps to select the position of the receiver with high perception. The fine-grained device-independent passive human body localization model based on S classification principle can improve the range and accuracy of indoor wireless device-independent passive localization and reduce the number of blind spots in the monitoring area. Finally, based on the difference and correlation of channel state information at different frequencies, a high-precision indoor device-independent passive location model is proposed. This paper makes full use of the frequency selective fading characteristic of channel response. Two novel passive human localization algorithms, weighted Bayesian localization and maximum similarity matrix localization, are studied under the condition of single link. Supervised learning technique is used to further improve the accuracy of passive human localization. Mean location error also reflects the stability and efficiency of channel response in device independent passive location.
【学位授予单位】：哈尔滨工程大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN92

【相似文献】