common destination: a physcial-layer device identification system

* classification
* verification

# Exploring the Physical-layer Identification of GSM Devices
## device
mobile devices(GSM devices) composed of different model and manufacturer and the same ones
## identification signal
transient and the data parts of the acquired GSM signals
## experiment:

* interaction between a GSM Device and a GSM basestation
* acquisition setup collecting signals
    1. data-related part: GSM Normal Burst including TS bits (26 training sequence bits)
	2. transient part: turn-on and turn-off transients
	![title](https://leanote.com/api/file/getImage?fileId=59f046cfab64414b9d001467)
* explored features
    ![title](https://leanote.com/api/file/getImage?fileId=59f04711ab6441495e001522)
    * turn-on and ture-off signal transients
        * the instantaneous phase $ \phi $
        * the instantaneous frequency $ f $
        * the signal power envelope $ e $
    ![title](https://leanote.com/api/file/getImage?fileId=59f04d75ab64414b9d0015b9)
    * the signal data part
        * phase and frequency errors
        * based on the spectral components of each bit of the training sequence.
        * based on 7 different statistical metrics computed over the phase (frequency) error of each bit in the training sequence.The considered metrics include mean μ, root mean square (RMS) r, peak p, standard deviation σ, variance σ2, skewness γ and kurtosis k.
    * compute the similarity score between a set of testing fingerprints and a set of reference fingerprints from all devices
    * use EER, where FAR(Flase Reject Rate) = FRR(False Accept Rate)
* 分类和筛选：欧氏距离
## result

## others

* sensitivity: 
    1. device position with respect to acquisition setup
	2. the device transmission power
* applicability: 
	1. local Swisscom network (uncontrol)
	2. an ad-hoc network based on the OpenBTS project. An open-source implementation of GSM protocal stack (controled parameters: carrier frequency, TS bits, device transimission power)

# Authorized and Rogue Device Discrimination Using Dimensionally Reduced RF-DNA Fingerprints
## signal
802.11a WiFi和802.16e WiMAX
![title](https://leanote.com/api/file/getImage?fileId=59f085acab64414b9d001d10)
![title](https://leanote.com/api/file/getImage?fileId=59f085bbab64414b9d001d11)
## 处理过程

### 特征提取
* Gabor 变换 Discrete Gabor Transform
    ![title](https://leanote.com/api/file/getImage?fileId=59f08b39ab6441495e001e1a)
    ![title](https://leanote.com/api/file/getImage?fileId=59f09122ab6441495e001f3f)
* 从归一化的小波系数矩阵中计算The T-F magnitude response
* statistics of standard deviation (σ), variance (σ2), skewness (γ ), and kurtosis (κ)

### 分类
GRLVQI分类器（？？？）
内置了一个DRA(Dimensional Reduction Analysis)
relevance ranking  top 10%

### 结果
#### 802.11a WiFi
![title](https://leanote.com/api/file/getImage?fileId=59f1a49fab644170ce001960)
#### wifi-max
![title](https://leanote.com/api/file/getImage?fileId=59f14dbdab64416e8b000ad7)
![title](https://leanote.com/api/file/getImage?fileId=59f14dd0ab644170ce000aa1)

逆风起笔