IEEE Signal Processing - May 2018 - 72

Table 2. The false alarm and detection probability for NLOS multievent
detection by the TRIEDS in a normal environment (location B) [62].
LOC B

Axis 1

Axis 2

Axis 3

Axis 4

TRIEDS detection rate (%)

96.92

98.95

99.23

99.4

TRIEDS false alarm rate (%)

3.08

1.05

0.77

0.6

RSSI detection rate (%)

92.5

94.16

94.77

95.36

RSSI false alarm rate (%)

7.5

5.84

5.23

4.64

Table 3. The false alarm and detection probability for LOS multievent
detection by the TRIEDS in a normal environment (location C) [62].
LOC C

Axis 1

Axis 2

Axis 3

Axis 4

TRIEDS detection rate (%)

97.89

98.94

99.18

99.36

TRIEDS false alarm rate (%)

2.11

1.06

0.82

0.64

RSSI detection rate (%)

96.73

97.19

97.35

97.43

RSSI false alarm rate (%)

3.27

2.81

2.65

2.57

FIGURE 16. A demonstration of radio propagation around a human body [77].

counterfeiting biometrics, techniques based on biological individualities are good at countering growing security threats and
facilitating personalization and convenience, possessing clearcut advantages over traditional security methods, such as passwords and signatures.
Although current biometrics systems are accurate and can
be applied in all environments, all of them require special
devices that capture human biometric traits in an extremely
LOS environment-i.e., the subject must make contact with
the devices. In this section, we discuss a novel concept of
radio biometrics utilizing the multipath CSI, based on which
accurate human identification and verification can be implemented with commercial Wi-Fi devices, even in a throughthe-wall setting.

CSI as human biometrics
The relationship between human physical characteristics and
the absorption of EM waves by human bodies has been studied in the carrier frequency range of 1-15 GHz, where Melia
found that the body's surface area has a dominant effect on
absorption [65]. Other researchers further studied the interaction of EM waves with biological tissues [66] and measured
72

these tissues' dielectric properties [67], [68]. According to the
literature, the wireless propagation pattern around a human
body depends highly on individual physical characteristics
(e.g., height and mass), the total body water volume, the skin
condition, and the characteristics of other biological tissues.
The work in [76] first proposed the concept of human radio
biometrics and defined it as the human-affected wireless signal under attenuations and alterations, containing individual
identity information. The term radio shot is defined as the
procedure of taking and recording human radio biometrics via
RF signals (such as Wi-Fi).
Radio biometrics can be viewed as unique for each human
being. Take the DNA sequence as an example. All humans are
99.5% similar in terms of DNA sequence, but no two are genetically identical, which is the key to techniques such as genetic
fingerprinting [69]. Considering the combination of all of the
physical characteristics and other biological features that affect the
propagation of EM waves around the human body and how variable those features can be among different individuals, the chance
for two humans to have identical combinations is significantly
small, no matter how similar those features might be. Consequently, a person's radio biometrics, which reflect how wireless
signals interact with the human body, can be viewed as unique.
The human body acts as an RF reflector and absorber in
an indoor wireless signal propagation environment. Radio
propagation around the human body is demonstrated in Figure 16, where the red dots represent the reflecting and scattering points due to the human body and other objects. As
seen from the figure, human radio biometrics are implicitly
embedded in the multipath CSI profile. However, the human
body may affect only a few paths of the multipath CSI, and the
energy of those paths is small because of the low reflectivity
and permittivity compared with other static objects, such as
the walls and furniture. As a result, human radio biometrics
captured through radio shot are buried in the CSI by other
useless components.
Furthermore, because the raw CSI obtained from Wi-Fi
chips is a large-scale, complex-valued matrix, the resulting
raw radio biometrics are high dimensional and complex valued, which further complicates the identification problem and
increases the computation complexity. To tackle the above
problem, we discuss several postprocessing algorithms to
calculate an analytic based on the TR technique that extracts
the human radio biometrics and emphasizes the differences
among individual radio biometrics.

TR-based human identification
TR-based human identification exploits the human radio biometrics embedded in the CSI and consists of the following
two key components:
■ Human radio biometrics refinement: This module extracts
the human biometric information from the raw CSI measurement, which is a complex-valued matrix. Because of
the independence of each link, the background for each
must be calculated and compensated for individually. An
important consideration is that each CSI measurement may

IEEE Signal Processing Magazine

May 2018

Table of Contents for the Digital Edition of IEEE Signal Processing - May 2018