IEEE Signal Processing - May 2018 - 60

This leads us to the conclusion that we might be able to extract meaningful information through wireless sensing by analyzing various features implanted in wireless signals.
By deploying wireless transceivers indoors, macro changes
due to human activities and moving objects can be derived
from the wireless signals. This process can help infer the
real-time location of a moving object [1]-[8], detect an event
[51]-[60], and facilitate applications in manufacturing asset
tracking, intelligent transportation, and home and office security systems. In addition, micro changes generated by gestures
[52] and vital signals [78]-[80] can be captured without requiring people to wear any device, which is especially useful for
providing assistance to the disabled and elderly people in smart
home applications.
The performance of wireless sensing depends greatly on
the richness of information that can be drawn from the radio
signals, while the information richness is often dictated by the
channel bandwidth through which the radio signals are transmitted. Because of the constrained bandwidth of the past, only
a limited number of multipaths could be seen, and not much
information could be revealed. But with more bandwidth available for the next generation of wireless systems, many more
smart IoT applications and services appear to be possible in
the near future, because richer information becomes available with a wider bandwidth. For example, with a much larger
bandwidth, one can see many more multipaths indoors, which
can serve as hundreds of virtual antennas and/or sensors ready
to assist in many applications.
The challenge here is how to control the virtual antennas to
meet our needs for smart IoT applications. We have to resort
to physics to do so, and the TR phenomenon is a good starting
point [37]. The TR technique treats each path of the multipath
channel as a distributed virtual antenna and provides a highresolution spatial-temporal resonance, commonly known as
the focusing effect [38]-[40]. In physics, the TR spatial-temporal resonance can be viewed as the result of the resonance of
the electromagnetic (EM) field in response to the environment
[41]. When the propagation environment changes, the involved
multipath signal varies correspondingly, and, consequently,
the spatial-temporal resonance also changes. Inspired by the
fundamental physical principle of TR, various types of analytics, referred to as radio analytics, that can decipher the radio
waves to reveal the activities around us, based on the wireless
channel state information (CSI), can be developed to enable
many cutting-edge IoT applications envisioned for a long time
but never achieved.
In this article, we will present the fundamental concept
of radio analytics and several types of this approach for IoT
applications, including indoor positioning, indoor tracking,
event detection, human radio biometrics, and vital signs monitoring. We will first discuss the impact of bandwidth on the
multipath CSI and the TR principles that can fully harvest the
multipath CSI. Then we discuss how to achieve centimeter
accuracy in wireless indoor positioning systems (IPSs) that
can offer an indoor Global Positioning System (GPS)-like
capability to track human or any indoor objects without any
60

infrastructure, as long as Wi-Fi or long-term evolution (LTE)
is available.
As such an IPS relies on CSI as a fingerprint and may require recalibration because of environmental changes, we then
discuss an indoor tracking system based on the statistical EM
property of the spatial focusing effect to achieve decimeter
accuracy without mapping and recalibration. Inspired by the
TR-based IPS that maps a location with a location-specific
CSI, we find that indoor wireless propagation depends highly
on the various indoor events and even the presence of different people. By mapping each event and/or person to the corresponding CSI, we can construct the notion of a TR space
in which one can quantify the similarity between two indoor
events and/or people and perform event detection and human
recognition. Since the multipath CSI is very sensitive to even
subtle environmental changes, we then discuss how to utilize
the CSI change pattern to detect vital signals. Finally, we will
survey and discuss recent related works.

Multipaths: A newfound friend
In wireless communications, when a signal emitted from a
transmitter is reflected or scattered by a scatterer, an attenuated copy of the original signal is generated and reaches the
receiver through a different path. The phenomenon that a signal is received by two or more paths is known as multipath
propagation. As depicted in Figure 1(a), where each scatterer is marked by a green star, the red arrow represents
the line-of-sight (LOS) path, while the blue arrows represent
paths reflected and scattered by scatterers. All of the paths
together form a multipath channel between the transmitter and
receiver [49]. Because two or more copies of the original signal arrive at the receiver and may be added in a noncoherent
way, multipaths can cause destructive interference and degrade
communication performance.
However, viewed from another perspective, the scatterers
in the environment act as virtual antennas and/or sensors that
can be leveraged to offer some desirable outcomes. Just imagine how everyday human activities involving motion and body
movements affect the wireless signal propagation around people and thus change the channel profiles and how information
about these activities becomes embedded in the signals. When
signals are bounced back and forth by the scatterers, multiple replicas are generated that contain enriched, meaningful
information about our activities. Each of these multipaths is,
in essence, a degree of freedom naturally existing in our surrounding environment. They can be considered to be tens or
hundreds of virtual antennas ready to serve us on demand.
To harvest multipaths, two key components to consider
are the transmission power and bandwidth [42], [43]. On the
one hand, increasing the transmission power leads to a higher
signal-to-noise ratio and thus to more observable multipath
components. On the other hand, the transmission bandwidth
determines the spatial resolution in resolving independent
multipath components. Because of the limited bandwidth,
which is equal to the channel sampling rate, the multipaths
with a propagation delay difference less than a channel sampling

IEEE Signal Processing Magazine

May 2018

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