WiFi多径环境下FTM发现精确测距法：MUSIC应用
需积分: 5 142 浏览量
更新于2024-08-06
收藏 5.49MB PDF 举报
"当FTM发现MUSIC：多径环境下精确的WiFi定位技术"
该论文标题《When FTM Discovered MUSIC》探讨了在无线网络环境中，特别是在存在多径效应的情况下，利用频域传输模态（Frequency-Time Modulation, FTM）实现精确WiFi定位的方法。WiFi信号的传播特性使得它容易受到多径效应的影响，这可能导致定位精度下降。MUSIC算法，全称为Multiple Signal Classification（多信号分类），是一种经典的信号处理技术，特别适用于估计信号源的方向和强度，从而在复杂的无线通信环境中进行高精度的位置估计。
作者Kevin Jiokeng、Gentian Jakllari、Alain Tchana和André-Luc Beylot的研究工作在2020年7月举行的IEEE International Conference on Computer Communications (INFOCOM)上发表，虽然是在虚拟会议上进行的，但其研究的重要性在于它提供了在多路径干扰条件下改善WiFi定位性能的新策略。他们提出的FTM与MUSIC结合的方法旨在克服传统WiFi定位中的难题，通过利用频域信息来区分不同的多径信号，从而提高定位的准确性。
论文详细地介绍了如何设计和实施这种基于频域传输的定位系统，可能包括信号采集、预处理、MUSIC算法的具体应用以及实验证据，展示了在理论分析和实际场景测试中，该方法的有效性和鲁棒性。此外，他们还提供了详细的性能评估指标，如定位误差和信噪比，以便其他研究人员可以理解和复制他们的工作，或者在此基础上进行进一步的优化和改进。
《When FTM Discovered MUSIC》是一篇具有实践价值的学术论文，对于那些关注室内定位、无线信号处理和通信系统优化的读者来说，它提供了一个创新且实用的解决方案，尤其是在多径环境下的WiFi定位挑战。通过阅读和理解这篇论文，研究者和技术人员能更好地理解和应用频域方法来提升无线定位系统的精度和可靠性。
have shown that it performs poorly on WiFi hardware.
In this work we present FUSIC, an approach fusing FTM
and MUSIC with the goal of extending FTM’s LOS accuracy
to NLOS settings. FUSIC requires no changes to the standard
– it simply takes as input the FTM ranging estimates, the WiFi
channel state information (CSI) readings, and corrects the error
when it detects that one has occurred. Thereby, FUSIC can be
implemented as a stand-alone, user-level application on mobile
devices and without requiring any modification to the access
points.
To realize its goal, FUSIC faces several challenges. First,
while the FTM performance in NLOS shown in Fig. 1 was
thoroughly evaluated in [13], two key underlying questions were
left open: 1) How do the multiple instances of the transmitted
signal and their relative strengths impact the FTM accuracy,
and 2) Do obstacles play an additional role in the observed
inaccuracy. The radio-wave signal slows down when it crosses
obstacles by a factor that depends on their relative permittivity.
FTM transforms the ToF to distance by using the speed of
light, leading to potential errors. The impact of such errors,
however, on FTM has not been studied yet.
Second, FUSIC needs to detect when the value returned by
FTM is inaccurate, even though this value and CSI are its only
input. A straightforward approach might be to apply MUSIC
on the CSI to obtain the power-delay profile and conclude there
is an error when the direct signal is not the strongest. However,
our measurements show that FTM can be accurate even when
the direct signal is not the strongest; applying correction on
an accurate result can make it erroneous. Finally, once FUSIC
detects that FTM has returned an inaccurate result it needs to
correct it while having as input only this value and MUSIC’s
power-delay profile – both erroneous.
In short, we address these challenges by first conducting a
measurement-based analysis on off-the-shelf hardware designed
to shed light on the factors leading to the poor performance
of WiFi FTM in NLOS. Coupling FTM output with MUSIC’s
power-delay profile of the received signals, we develop a fine-
grained understanding of the relation between multipath and
FTM. We leverage this understanding for designing FUSIC’s
two integral parts: a mechanism for identifying when FTM
returns erroneous results, and an error correction mechanism
fusing data from FTM and MUSIC.
Our main contributions may be summarized as follows:
•
In Section III, we conduct a measurement analysis of WiFi
FTM and MUSIC using off-the-shelf hardware. We assess
the magnitude of the FTM weaknesses in NLOS and study
the underlying reasons at the signal level. Furthermore, we
assess MUSIC’s ability to help improve the accuracy of
FTM.
•
In Section IV, we use the lessons learned in our measurement
analysis to design FUSIC, an algorithm that takes as input
the FTM distance estimates and CSI and 1) is able to detect
whether the FTM ranging result is inaccurate and 2) correct
the FTM ranging result when necessary.
•
In Section V, we use a testbed comprising off-the-shelf
hardware to conduct an extensive evaluation of FUSIC in
4 different physical settings, including a controlled setting,
a university restaurant, a warehouse and a student lounge.
Our experiments show that a) FUSIC extends FTM’s LOS
ranging accuracy to NLOS settings – hence, achieving its
stated goal; b) it significantly improves FTM’s capability to
offer room-level indoor positioning.
II. BACKGROUND
This section presents the necessary background to understand
our contribution.
A. Channel State Information (CSI)
In wireless systems, the signal that reaches a receiver is
generally altered (eg. attenuated and reflected) by the channel
in which it travels before reaching the receiver. If we denote by
x
the signal sent by the transmitter, the signal
y
that reaches
the receiver is given by the equation
y = H ∗ x + n (1)
where the matrix
H
represents the complex attenuation and
phase shifts undergone by the signal while going through
the channel, and
n
the ambient noise, often assumed to be
white Gaussian with zero mean.
H
is called Channel State
Information (CSI) and represents the properties of the channel
between the sender and the receiver. Many research works [6],
[7], [15]–[21] have used CSI in their solutions as processing
them can give useful information about the signal propagation,
including Time of Flight (ToF), Angle of Arrival (AoA) and
Power Delay Profile (PDP).
B. MUSIC in the frequency domain: ToF estimation of multiple
propagation paths
In the interest of brevity, we provide an intuitive summary
of MUSIC (MUltiple SIgnal Classification) [14], necessary
to understand our work. For an in-depth description we refer
the interested reader to [14], [15], [18]. MUSIC algorithm
distinguishes signals based on predictable variations of phase
when it comes from a specific location. It relies on the
measurements obtained from each subcarrier of an OFDM
system (as is the case of WiFi, for example). This is feasible
because for a given ToF, a difference in terms of signal
frequency produces a difference in terms of phase at the
receiving system. In fact, two signals that reach an antenna
after having travelled during a ToF
τ
will reach that antenna
with a predictable phase difference of
−2π × (f
j
− f
i
) × τ
,
with
f
j
and
f
i
being the frequencies of those signals. This
means that, with the knowledge of signal measurements on
different subcarriers of an OFDM WiFi band, it is possible
to resolve the ToFs over different propagation paths. MUSIC
uses this property to build a model that is able to resolve
the ToFs of different propagation paths. MUSIC algorithm
takes as input the CSI corresponding to the communication
and returns a spectrum indicating the signal power perceived
at each instant by the receiver, a kind of PDP. From such a
spectrum, propagation paths can be identified by taking the
peaks of the spectrum. This gives an estimate of their ToFs
and Power.
剩余10页未读，继续阅读
2022-09-21 上传
2020-12-30 上传
2021-09-25 上传
2021-09-30 上传
2020-02-17 上传
2023-02-13 上传
2019-10-19 上传
2021-11-18 上传
2023-02-13 上传
RenMinSherry
- 粉丝: 0
- 资源: 5
我的内容管理
展开
最新资源
- NIST REFPROP问题反馈与解决方案存储库
- 掌握LeetCode习题的系统开源答案
- ctop：实现汉字按首字母拼音分类排序的PHP工具
- 微信小程序课程学习——投资融资类产品说明
- Matlab犯罪模拟器开发：探索《当蛮力失败》犯罪惩罚模型
- Java网上招聘系统实战项目源码及部署教程
- OneSky APIPHP5库：PHP5.1及以上版本的API集成
- 实时监控MySQL导入进度的bash脚本技巧
- 使用MATLAB开发交流电压脉冲生成控制系统
- ESP32安全OTA更新：原生API与WebSocket加密传输
- Sonic-Sharp: 基于《刺猬索尼克》的开源C#游戏引擎
- Java文章发布系统源码及部署教程
- CQUPT Python课程代码资源完整分享
- 易语言实现获取目录尺寸的Scripting.FileSystemObject对象方法
- Excel宾果卡生成器：自定义和打印多张卡片
- 使用HALCON实现图像二维码自动读取与解码
