Ambient Noise Analysis on Sound For Use in Wireless Digital Transmission

Evaluation of Sound Perception to IdentifyCandidate Frequency for Wireless Networking

Kuruvilla Mathew, Chong Eng Tan, and Biju Issac

Abstract

Wireless technology has been introduced and growing since early twentieth century, but

there are still environments the current technologies find it difficult to penetrate. The dense

jungle terrain, for example, pose a huge challenge for the 0.12 m wavelength of the Wi-Fi

signals, but the FM radio frequency signals at a wavelength of 3 m function a lot better.

This paper studies the possibility of using a very low frequency, down to the range of

audible frequencies to try and identify the frequency band that can be used, ubiquitously

and unobtrusively. Sound can be considered as a ubiquitous signal due to obvious reasons

and the search is to find the unobtrusive frequency band that can be a candidate frequency

for data carrier signals. The paper is presented in two sections, the first section does a

geographically and age neutral survey to identify the unobtrusive signal and second section

analyses the noise profiles in these frequency bands.

Keywords

Ubiquitous computing � Rural networking � Low frequency network signal � Wireless

networking

Introduction

As the world is moving closer to bringing network technol-

ogy closer to the people, the need for ubiquitous mode of

operation is now highly pronounced. The Ubiquitous wire-

less networks need more than the current Wi-Fi signal

architectures in order to be more power efficient, better

performance in obstructions and for application varied

types of environments. One area in which the Wi-Fi in the

2.4 GHz band and other high frequency signals used for

communication under the FCC regulations fall short in

delivering efficient connectivity is in dense jungle type of

environments [10]. Where high frequency signals fail to

perform in environments with obstructions, research for

low frequency signals to deliver the required connectivity

begins. Low frequency signals are not only expected to

perform better in the presence of obstacles due to their

longer wavelength, they also need much lower power to

generate, and hence more sustainable in domains with lim-

ited power availability. However, it is also to be noted that as

the frequency becomes lower, the maximum bit-rate that can

be encoded is also lower and hence usually suitable for low

bandwidth connections. It is also advantageous to be able to

use some ubiquitous signal, which allows use of existing

devices with little or no change for it implementation. Such a

system is expected to have minimum cost impact, minimal

or no training requirements (for the new system) and a very

high acceptance factor. In this light, we are studying the

K. Mathew (*)

FECS, Swinburne University of Technology (Sarawak Campus),

Kuching, Sarawak, Malaysia

e-mail: [email protected]

C.E. Tan

F-IT & CS, University Malaysia Sarawak (UNIMAS),

Kuching, Malaysia


B. Issac

School of Computing, Teesside University, Middlesbrough, UK


K. Elleithy and T. Sobh (eds.), New Trends in Networking, Computing, E-learning,Systems Sciences, and Engineering, Lecture Notes in Electrical Engineering 312,

DOI 10.1007/978-3-319-06764-3_43, # Springer International Publishing Switzerland 2015

349

mailto:[email protected]



possibility of using a very low frequency, ubiquitous signal,

which is “sound”.

Sound can be generated and sensed with very cheap

hardware and is known to travel quite well around obstacles.

However, sound being an audible signal, and hence

perceivable to the human ear, can cause unwanted

distractions or disturbances its environment of use. Hence

the acceptance factor will be higher if sound frequencies

outside the hearing range for humans are used. The accepted

norm for hearing range in humans is from 20 to 20,000 Hz,

but the hearing threshold in humans is normally not tested

above 8 kHz [9]. We would therefore do random survey to

try and measure the practical audible frequency range

humans can “perceive”. We have conducted this as a cultur-

ally, demographically and age neutral study and noticed

that the audible frequency band is actually lower than

the accepted norm. The participants of the study was

from India, Indonesia, Malaysia, Australia, Netherlands,

Hungary, Korea, Brazil, Japan, Switzerland and England

with an age group ranging from 15 to 55 years. We will

use this output to identify the candidate frequency band that

we can use for the network architecture using sound as the

carrier signal. We will then conduct an analysis of the

ambient noise profiles in these frequency bands and compare

them as an approach to looking at the practical implications

of making use of these frequency bands.

Related Work

K.MathewandB. Issac presented sound as candidate carrier for

low bandwidth, low power communication, ubiquitously [1].

The paper demonstrated as proof of concept, the ubiquitous

data communicationusing consumer hardware in smart devices

and sound as the carrier signal.

K. Mathew, C.E. Tan and B. Issac studied the ambient

noise in various natural environments [2] in order to identify

the candidate frequency band for communication. The study

compares the noise profiles in common environments.

M. Weiser, presented pervasive computing technologies

that disappear into everyday life as they ubiquitously blend

into our daily activities so that they are indistinguishable [3].

The paper introduced concepts of tabs, pads and boards, and

opened up some challenge in networking which the nature of

the devices will present.

Madhavapeddy, Scott and Tseworked worked on audio

networking as a forgotten technology [4]. They successfully

sent and received data using sound as the carrier signal, with

common computing platforms to use high frequency audio

(ultrasonic) for the communication.

Chen and Lee looked at the inter-relationships among

major research themes in ubiquitous paradigm as a biblio-

graphic study on Ubiquitous Computing [5].

Jurdak, Lopes and Baldi proposed using acoustic signals

to uniquely identify and locate a user in an acoustic identifi-

cation scheme for location systems [6].

Madhavapeddy, Scott and Sharp presented context aware

computing with sound [7], analysing some location aware

applications including pickup and drop interface, digital

attachments in voice, etc.

Mandalet et al. proposed application of indoor position-

ing with 3D multilateration algorithms using audible sound

[8] and was able to give accuracy to about 2 feet almost 97 %

times using cheap consumer hardware.

A.R. Moller provides detailed information about the

physiology and anatomy of the entire auditory system in

humans as it describes disorders in hearing in his book

Hearing: Anatomy, Physiology and Disorders of the Audi-

tory System [9].

Popleteev, A., Osmani, V. and Mayora, O. presented an

investigation of indoor localization with ambient FM radio

stations [10], exploring the performance of Wi-Fi and other

signals in the indoor vs. outdoor scenarios.

Theory and Background

As computer networks grew, the users also became more

and more mobile and wireless networks came into being.

Wireless technologies makes use of various frequencies

in the radio frequency spectrum, ranging from low frequency

signals as low as a few kHz, the TV broadcast signals in the,

the more recent mobile phone networks, the very noisy,

unlicensed 2.4 GHz band used by a number of short distance

wireless protocols including Wi-Fi, Bluetooth, Zigbee and

many more in the IEEE 802.11(x) specifications, and some

ultra (UHF), extended (EHF) and Super (SHF) high

frequencies. The details of these specifications are found in

the Federal Communications Commission (FCC) online

table of frequency allocations [11]. A quick view of the

popular wireless bands in use in the electromagnetic spec-

trum is shown in Fig. 1. Each of the bands works in their

specific domains for providing specific services.

Popular Wireless Signals

The Wi-Fi is a short range signal operating within a room

or a small building. Mobile operators establish large towers

called base stations to cover a much larger areas and

mobile devices connect with these to establish communi-

cation Bluetooth was introduced as a cable replacement

standard and works in very short ranges, most often to

connect peripheral devices like headphones, microphones,

keyboards etc. to computing devices. The Zigbee related

350 K. Mathew et al.

protocols operate in under the IEEE 802.115 standards to

bring wireless network connectivity to low powered

devices. Performance of all of these assumes normal

urban environment and have seen to fall short in rural

environments with thick foliage. This could mainly be

due to the fact that the hilly and jungle terrain pose many

obstacles in the form of hills and valleys, trees, bushes,

rocks and many unpredictable entities. This calls for the

need for alternate architecture for effective networking in

the rural environments.

The Nature of the Signals

The popular signals we see so far in use for wireless data

communications, namely Wi-Fi, mobile, WiMax, etc. use

very high frequency, above the 800 MHz range. The high

frequency signals require more power to generate, but can

support higher speeds of data over greater distances. How-

ever, high frequency signals have shorter wavelength and

hence do not travel very well around obstacles causing the

range of coverage to drop drastically in such environments.

The Terrain

The popular wireless signal performs well in the urban

scenarios for which they are designed for, and we have

evidenced over the past that the speeds and Quality of

Services have kept on increasing with time. These are quite

inefficient in rural setting with lots of obstacles. The jungle

with thick foliage, with bushes in particular, is a proven

signal killer and is known to reduce the usable range by

about 25 %.

The high frequency signal of Wi-Fi, operating at a

frequency of 2.4 GHz has a wavelength of 0.12 m, and is

affected even by moving leaves [10]. This type of terrain

calls for the need of a different kind of signal and an archi-

tecture to make communication possible in remote terrains

which may be limited not just by the obstacles, but also by

the limited availability of power.

The New Signal Idea

An approach to tackle the above problem is to make use of a

low power, longer wavelength signal (which directly

translates to lower frequency) and low cost signal that can

function over the terrains in consideration. Lower frequency

implies a longer wavelength, which translates to better tra-

versal around obstacles. The FM frequency, operating at

about 100 MHz has an effective wavelength of about 3 m

and therefore travels well around most naturally found

obstructions like leaves and bushes. Hence the idea to

study the use of much lower acoustic frequencies arises.

Acoustic frequencies are the range of frequencies audible

to the human ear, and this is commonly seen as from 20 to

20,000 Hz. To make the communication effective, we will

need to use an acoustic signal above the ambient noise

levels. This however creates additional human perceivable

noise in the environment and hence may not be well

accepted. We will therefore study the range of frequencies

normally perceivable by people using a random sample

study as it is possible that the actual perceivable frequency

range is much shorter than the theoretical maximum range. If

this is possible, then this translates to the fact that there are

some frequencies in the acoustic spectrum that we can utilize

for data transmission.

Fig. 1 The electromagnetic

spectrum

Evaluation of Sound Perception to Identify Candidate Frequency for Wireless Networking 351

The Survey

A study is to be carried out among the general population to

try and identify the perceivable range of audible frequencies.

This study can be used to get an idea of the frequencies that

is outside the perceivable range and can be used as commu-

nication signals. We have the option of utilizing subsonic

and supersonic frequencies. The subsonic frequencies are

expected to be closer to the 20 Hz mark and hence should

be able to traverse well across obstructions. The supersonic

frequencies, closer to the 20 kHz mark, should theoretically

be able to carry more data as it has more signals per second.

The survey will therefore ask the participants to listen to pre-

generated signals of various frequencies and identify which

of these they can hear and which they cannot. This will help

us get to a range of frequencies that practically falls into the

audible space.

The Survey, Analysis and Results

The experiment has two phases, the first phase carried out a

survey with some generated signals and specific hardware

to evaluate the perception of the signals among the crowd.

The second phase evaluates the spectrum outside the audible

frequency range and performs a noise analysis on this.

This range is our candidate frequency band for our

communication.

The Sound Signals for the Survey

Phase 1 of the experiment involves a survey. In the survey,

we have generated sound samples of specific frequencies,

namely, 15, 20, 25, 30, 35, 40, 50, 100, 150 and 250 Hz in the

low frequency spectrum and 10, 12, 14, 16, 17, 18, 19, 20, 21

and 22 kHz. The signals were generated using the free open

source application “Audacity” and the frequency of the

sound signals generated was verified by plotting the fre-

quency of each signal on Matlab to notice that the generated

signal pattern is a sinusoidal wave of the correct peak

frequency.

The Survey Equipment and Hardware

The Survey involves playing back frequencies ranging from

20 Hz to 20 kHz. This implies that the audio hardware

should be capable of handling this wide range of

frequencies, which cannot be expected from usual consumer

hardware. We used the following hardware to maintain the

frequency range is reproduced as close as possible to natural,

uncoloured sound. Figure 2 shows the hardware used in the

survey.

1. AKG K99 Studio Monitor Headphones

The AKG K99 Studio Monitor headphones offer a fre-

quency response range from 18 to 22,000 Hz. They pro-

vide natural, uncoloured sound which is a requirement for

the validity of this survey.

2. Audio Player (Apple iPad)

We needed any audio player that will reproduce the

digital audio signals we have generated without any

colouring. We used the Apple iPad to play out the

sound through the AKG headphones for the experiment.

3. Other Devices for ComparisonWe also tried a comparison using other smart devices and

with popular consumer earphones to notice the variation

in the participant response and noticed the colouring the

consumer devices adds to the original sound signals. The

result of this comparison is not included as it does not

directly contribute to this study. Some of the devices were

Samsung and HTC Smartphones along with their bundled

earphones, an Asus laptop with standard headphones etc.

We also noticed some playback software can also cause

or add colouring to the signal, which therefore was care-

fully avoided before the survey.

The Survey Process

The survey was done in normal user environment, which can

be considered to be moderate to low noise environments.

The participants were first asked to listen to a white noise in

to set a comfortable listening volume. Once this is set, the

signals are played back one by one, from the lowest to higher

Fig. 2 The hardware used for the experiment and survey. (1) Apple

iPad used to play back the signals, (2) AKG K99, Zoom H4n with

remote, consumer headphones (Samsung and HTC)


frequency for the low frequency band and from the highest

to lower frequency for the high frequency band. The very

low frequencies are not audible (15 Hz) and the participant

will respond to the first audible signal, which is recorded as

the lowest frequency the person can “hear” or “perceive as

present”. Once the lowest perceivable frequency is

identified, we start from the very high frequency, at 22

kHz, which also is not normally audible, and play back the

lower frequency signals, one at a time. The participant will

respond when they can hear the audio signal, which is

recorded as the highest frequency that they can hear. In the

case of high frequency signal, we expect that they can

actually hear and not “feel” the sound as in the case of low

frequency.

The sample selected for the survey included participants

from all over the world, from the Americas, Africa, Europe,

Asia and the Oceania/Australia and hence is culturally and

geographically neutral. The population is also age neutral

and includes participants from various age groups within 15

years and 55 years, which helps the survey results not to be

biased towards a particular age group. The identities of the

participants are kept confidential in order to protect their

privacy.

Survey Results

Table 1 shows the results of this study, where we notice that

the lowest frequency of perceivable audio signal observed in

the study was 25 Hz and the highest was 18 kHz. This allows

us to deduce that majority of the population may not per-

ceive the presence of a sound signal below 30 Hz, at normal

amplitudes, though at a really loud amplitude we may “feel”

the very low frequency vibrations [10]. The distribution of

participants’ response to the perceivable low frequency sig-

nal is shown on Fig. 3 and the perceivable high frequency

signal is shown on Fig. 4. We can observe that from the

population, no one could hear the tones of 15 and 20 Hz.

Though some could “perceive” the presence of the 25 Hz

signal and majority could perceive only from 35 Hz and

above. On the high frequency side, though we can notice

that a few can “perceive” or actually “hear” the audio signal

up to 18 kHz, majority cannot perceive above 14 kHz, and

some can perceive up to 17 kHz. 18 kHz, can therefore, be

considered a safe upper limit of “perceivable audible sound

frequency” according to the results of this survey.

Noise Analysis of Candidate Frequency Band

The survey has helped us narrow down on candidate fre-

quency band that can be considered available for “audio

networking”. However, these being sound signals, is still

open to a lot of ambient noise that may be present in the

natural environments of intended use. Hence we also car-

ried out a noise profiling for the candidate frequency

bands.

Table 1 Results of audio perception survey

DocN Geo AG Low (Hz) High (kHz)

JL1325-01 India, Middle/ S. Asia <15 25 17

JL1325-02 Indonesia, E. Asia <30 30 16

JL1325-03 Malaysia, E. Asia <30 50 17

JL1325-04 Australia, Oceania <30 35 17

JL1325-05 Netherlands, Europe <30 35 16

JL1325-06 Hungary, Europe <30 35 14

JL1325-07 Korea, E. Asia <55 50 14

JL1326-01 Brazil, S. America <30 30 14


JL1329-01 Japan, E. Asia <55 25 14

JL1329-02 Switzerland, Europe <30 30 14


JL1330-02 England, Europe <30 30 17

DocN, participant response document number (name not included to

protect respondent privacy); Geo, geography/country; AG, age group;

Low, lowest perceived frequency; High, highest perceived frequency

Fig. 3 Survey result plotted for frequency of sound against number of

participants who’s lowest perceivable frequency

Fig. 4 Survey result plotted for frequency of sound against number of

participants whose highest perceivable frequency


The Candidate Frequency Bands

As a result of the study, we have identified that the frequency

bands below 25 Hz and above 18 kHz can be considered for

the intended audio networking technology. The noise

profiling was done in the four environments we studied,

namely, the quiet office, the busy cafeteria, the beach and

the urban roadside. For the purpose of comparison, we have

divided the zones into the low frequency zone and the high

frequency zone. The further sections describe the analysis

and the results of the studies.

The Experiment Process

This analysis was conducted by recording the ambient noise

from various environments using the Zoom H4n recorder,

shown in Fig. 2. All the sound samples were recorded with

the same device, using its high response built in

microphones at a consistent gain of 40 % in order to get

comparable signal. The recorded signals were analysed

using Matlab to plot the actual spectrum of the noise and

then specific bands we are interested in using a band (low

pass and high pass) filter and plotted within 0 to 10�5 db.

The results of the analysis are discussed in the next sections.

The Low Frequency Zone

The very low audio frequencies outside the perceivable

range noticed are below the 25 Hz mark. The very high

audio frequencies outside the perceivable range noticed are

above the 18 kHz mark. The study of the comparison of

noise profiles in “The Cafeteria”, “Car Park”, “City Side

Park”, “Beach at Night with Footsteps” and “The Crowded

Lobby” is shown in Figs. 5, 6, 7, 8 and 9 respectively.

Full Range Power Spectrum of Sound: 01-CafeteriaAudioRecording.wav

Power Spectrum of Low Freq. Band of the Sound: 01-CafeteriaAudioRecording.wav

Power Spectrum of High Freq. Band of the Sound: 01-CafeteriaAudioRecording.wav

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–61.7

0

0 5

0.2 0.4 0.6 0.8Frequency, Hz

Frequency, Hz

Frequency, Hz

1 1.2 1.4 1.6 1.8 2

1.75 1.8 1.85 1.9 1.95 2x 104

x 104

10 15 20 25 30

Fig. 5 Cafeteria noise analysis


Few more environments were analysed, but not included as

it did not add more light to the results.

We can notice a very consistent pattern in all the

signals recorded and analysed. We can see that in almost

all cases, we have strong low frequency signal in the

range of 10�5 db to 10�2 db, but the high frequencies

above 18 kHz is almost absent. The high frequency signal,

on the other hand, has a much larger bandwidth above 18

kHz and is theoretically expected to be able to support

better data rates. Hence, considering the noise to signal

ratio and possible future performance requirements, the

high frequency band above 18 kHz is a better

recommended candidate for the intended audio network-

ing architecture.

Full Range Power Spectrum of Sound: 03-CarParkingLot.wav

Power Spectrum of Low Freq. Band of the Sound: 03-CarParkingLot.wav

Power Spectrum of High Freq. Band of the Sound: 03-CarParkingLot.wav

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–61.7

0

0 5

0.2 0.4 0.6 0.8Frequency, Hz

Frequency, Hz

Frequency, Hz

1 1.2 1.4 1.6 1.8 2

1.75 1.8 1.85 1.9 1.95 2× 104

× 104

10 15 20 25 30

Fig. 6 The car park noise analysis


Conclusion

The study was aimed and has contributed towards

identifying a candidate sound frequency for signal trans-

mission, for use in network technology, unobtrusively.

Using of the frequencies outside the perceivable range

will allow network communication to carry on using

sound signals that blend or fade away into the background.

In an effort to further the study, we have also done a noise

sampling for a number of environments and did a profiling

on the identified candidate frequency bands to notice that

ambient noise in the high frequency band is a lot “quieter”

that the low frequency band, suggesting that as a better

possible channel for network communication. The avail-

able frequency range in this band is a lot wider as well as

having a much higher signal rate, suggesting possibility of

a theoretical higher data rate. It has also been noticed that

Full Range Power Spectrum of Sound: 04-CitySidePark.wav

Power Spectrum of Low Freq. Band of the Sound: 04-CitySidePark.wav

Power Spectrum of High Freq. Band of the Sound: 04-CitySidePark.wav

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–61.7

0

0 5

0.2 0.4 0.6 0.8Frequency, Hz

Frequency, Hz

Frequency, Hz

1 1.2 1.4 1.6 1.8 2

1.75 1.8 1.85 1.9 1.95 2× 104

× 104

10 15 20 25 30

Fig. 7 City side park noise analysis


very low frequency sounds played at very high amplitude,

even though may not heard, can be “felt” or “perceived” by

the humans [11]. Further studies need to be carried out to

identify ubiquitous devices that can handle the frequencies

well. The actual attenuation of the various signal

frequencies in different environments also need to be stud-

ied further to evaluate sufficiency in terms of signal propa-

gation in our target environments.

Full Range Power Spectrum of Sound: 06-BeachSide-FootstepsAtNight.wav

Power Spectrum of Low Freq. Band of the Sound: 06-BeachSide-FootstepsAtNight.wav

Power Spectrum of High Freq. Band of the Sound: 06-BeachSide-FootstepsAtNight.wav

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–61.7

0

0 5

0.2 0.4 0.6 0.8Frequency, Hz

Frequency, Hz

Frequency, Hz

1 1.2 1.4 1.6 1.8 2

1.75 1.8 1.85 1.9 1.95 2x 104

x 104

10 15 20 25 30

Fig. 8 The beach at night with footsteps noise analysis


Full Range Power Spectrum of Sound: 02-CrowdedLobby.wav

Power Spectrum of Low Freq. Band of the Sound: 02-CrowdedLobby.wav

Power Spectrum of High Freq. Band of the Sound: 02-CrowdedLobby.wav

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

100

10–2

Pow

er a

t fre

quen

cy, d

B

10–4

10–6

1.7

0

0 5

0.2 0.4 0.6 0.8Frequency, Hz

Frequency, Hz

Frequency, Hz

1 1.2 1.4 1.6 1.8 2

1.75 1.8 1.85 1.9 1.95 2× 104

× 104

10 15 20 25 30

Fig. 9 The crowded lobby noise analysis


References

1. K. Mathew and B. Issac“Ubiquitous text transfer using sound -

Azero-infrastructure alternative for simple text communication”.

Proceedings of the International Joint Conferences on Computer,

Information, and Systems Sciences, and Engineering – Interna-

tional Conference on Telecommunications and Networking (TeNe

2012), USA, Springer.

2. K. Mathew, B. Issac, C.E. Tan, “Ambient noise analysis on sound

for use in wireless digital transmission,” Information Technology in

Asia (CITA), 2013 8th International Conference on , vol., no.,

pp.1,4, 1-4 July 2013. doi: 10.1109/CITA.2013.6637559

3. M. Weiser, “The computer for the 21st Century”. IEEE Pervasive

Computing, 99(1), 19–25.

4. A. Madhavapeddy, R. Sharp, D. Scott and A. Tse, “Audio network-

ing: the forgotten wireless technology”, Pervasive Computing,

IEEE , vol.4, no.3, pp. 55–60, 2005.

5. T. T.Chen and M. Lee, “Ubiquitous computing in prospect: A

bibliographic study”. International Symposium on Ubiquitous Mul-

timedia Computing, 2008, pp.57–62.

6. R. Jurdak, C. V. Lopes, and P. Baldi, (n.d.). “An acoustic identifi-

cation scheme for location systems”, Proceedings of IEEE/ACS

International Conference on Pervasive Services, 2004, pp.61–70.

7. A. Madhavapeddy, D. Scott, and R. Sharp, “Context-aware com-

puting with sound”, In the proceedings of The Fifth International

Conference on Ubiquitous Computing, 2003, pp.315–332.

8. A. Mandal, C. V. Lopes, T. Givargis, A. Haghighat, R. Jurdak,

and P. Baldi, “Beep: 3D indoor positioning using audible sound”,

Second IEEE Conference on Consumer Communications and Net-

working, 2005, pp. 348–353.

9. Moller A.R Hearing: anatomy, physiology and disorders of the

auditory systems. In Academic Press 2nd edn. 2006 London, UK:

Academic Press

10. Popleteev, A., Osmani, V., & Mayora, O. (2012, March). Investi-

gation of indoor localization with ambient FM radio stations. In

Pervasive Computing and Communications (PerCom), 2012 IEEE

International Conference on (pp. 171–179). IEEE.

11. “FCC - Frequency Allocation Table.” OET - Frequency Allocation

Table. Federal Communications Commission Office of Engieneering

and Technology PPolicies and Rules, n.d. Web. 01 Nov. 2013.

<http://www.fcc.gov/oet/spectrum/table/Welcome.html>.


http://dx.doi.org/10.1109/CITA.2013.6637559

http://www.fcc.gov/oet/spectrum/table/Welcome.html

Ambient Noise Analysis on Sound For Use in Wireless Digital Transmission

Documents

Transcript of Ambient Noise Analysis on Sound For Use in Wireless Digital Transmission