Signal Processing - September 2017 - 130

Table 2. The teams that participated in the 2016 and 2017 Microsoft Indoor Localization Competitions. Teams in each category
are listed in order of the localization accuracy they achieved (highest to lowest) that year. (continued )
Team

Team's Affiliation

Country

Technical Approach

Global
Rank

RealEarth (init.) et al. [79]

Kaarta

United States

Lidar

1

Quantitec Intranav [66]

Quantitec Intranav

Germany

UWB

2

Beuchat et al. [98]

ETH

Switzerland

UWB

3

Chen et al. [69]

King Abdullah University of
Science and Technology

Saudi Arabia

Sound

6

Cheng and Guan [99]

Kiwii Power Technology
Corporation

United States

UWB

7

Snyder et al. (init.) [100]

Astrobotic

United States

Lidar

10

Wang et al. [101]

McMaster University

Canada

Sound

11

Acton and Kulkarni [102]

Arin Technologies

United States

UWB

12

Hemamali and Sreenivas [103]

Trackray

India

UWB

13

Li et al. [104]

McMaster University

Canada

IMU

4

Ju (init.) et al. [105]

Seoul National University

South Korea

Foot-mounted IMU

5

Kikutchi et al. [106]

Iwate University

Japan

Direction of departure-based BLE
localization

8

Su et al. [107]

National Taiwan University of
Science and Technology

Taiwan

Wi-Fi + magnetic + IMU
fingerprinting

9

Ben-Moshe et al. [108]

Ariel University

Israel

Camera and range sensing

14

Sadhu et al. [109]

Rutgers University

United States

Wi-Fi fingerprinting

15

Guimarães et al. [39]

Fraunhofer Research Institute

Portugal

Wi-Fi + magnetic + IMU
fingerprinting

16

2017
Infrastructure based

Infrastructure free

wireless signal travels is known, the distance between the two
devices can be easily extracted.
Alternatively, time difference of flight (TDoF) techniques
can be used to estimate the distance between the devices. One
of the devices can transmit two signals of different types at
the same time (i.e., an ultrasound signal and an RF signal).
Observe that these two signals travel at different speeds; RF
travels at the speed of light, while ultrasound signals travel at
the speed of sound. The receiving device can timestamp when
these signals are received, and it can then use the equation in
Figure 2 to estimate the distance between the Tx and Rx.
In some cases, the distance between two devices can be measured using a third transmitting device, as shown in Figure  2.
Instead of measuring the TDoF, we can now measure the TDoA
of two wireless transmissions. Device Tx(1) in Figure 2 makes a
single wireless transmission. As soon as device Tx(2) receives
the transmission from device Tx(1), it initiates its own transmission. Device Rx timestamps the reception of both transmissions and computes the distance between devices Tx(1)
and Tx(2) according to the equation in Figure 2.

Wireless mediums for ranging
The type of signals used for the transmissions shown in
Figure  2 is critical. The most popular wireless signals for
ranging are RF and sound/ultrasound signals. RF signals have
130

excellent penetration characteristics, which make them ideal
for indoor localization scenarios. In addition, there is already
widely deployed RF infrastructure, such as Wi-Fi access
points, that can be leveraged for localization purposes. On the
negative side, RF signals travel at the speed of light, which
makes accurate timestamping a challenging problem to solve.
Inaccurate timestamps can lead to large ranging errors and,
therefore, to poor localization accuracy.
Sound and ultrasound signals [26], [27] have the advantage of lower propagation speed over RF signals. The lower
the propagation speed of the wireless transmission, the easier
it is to design the necessary hardware for accurately timestamping these transmissions. On the negative side, sound and
ultrasound signals require line of sight, as they do not penetrate materials like walls. In addition, this type of signal can
interfere with hearing, as they are audible by humans (sound
signals) and animals (ultrasound signals).
More recently, UWB signals [40] have become increasingly
popular for indoor localization applications. UWB signals are
unique, as they consist of very narrow pulses transmitted over a
large bandwidth, usually in the order of gighertz. Even though
they overlap with multiple frequencies, no interference is created, as the transmission power at each frequency is very low.
UWB transmissions require the Tx and Rx to be coordinated to
send and receive pulses. The advantage of UWB transmissions

IEEE SIGNAL PROCESSING MAGAZINE

|

September 2017

|
Table of Contents for the Digital Edition of Signal Processing - September 2017
