Towards Automated Smart Mobile Crowdsensing for Tinnitus Research [original]

Towards Automated Smart Mobile Crowdsensing

for Tinnitus Research

Muntazir Mehdi∗, Denis Schwager∗, Rüdiger Pryss†, Winfried Schlee‡, Manfred Reichert†, Franz J. Hauck∗

∗Institute of Distributed Systems, Ulm University

†Institute of Databases and Information Systems, Ulm University

Email: {muntazir.mehdi,denis-norbert.schwager,ruediger.pryss,manfred.reichert,franz.hauck}@uni-ulm.de

‡Clinic and Policlinic for Psychiatry and Psychotherapy, Regensburg

Email: [email protected]g

Abstract—Tinnitus is a disorder that is not entirely understood,

and many of its correlations are still unknown. On the other

hand, smartphones became ubiquitous. Their modern versions

provide high computational capabilities, reasonable battery size,

and a bunch of embedded high-quality sensors, combined with

an accepted user interface and an application ecosystem. For

tinnitus, as for many other health problems, there are a number

of apps trying to help patients, therapists, and researchers to

get insights into personal characteristics but also into scientific

correlations as such. In this paper, we present the first approach

to an app in this context, called TinnituSense that does automatic

sensing of related characteristics and enables correlations to the

current condition of the patient by a combined participatory

sensing, e.g., a questionnaire. For tinnitus, there is a strong

hypothesis that weather conditions have some influence. Our

proof-of-concept implementation records weather-related sensor

data and correlates them to the standard Tinnitus Handicap

Inventory (THI) questionnaire. Thus, TinnituSense enables ther-

apists and researchers to collect evidence for unknown facts,

as this is the first opportunity to correlate weather to patient

conditions on a larger scale. Our concept as such is limited neither

to tinnitus nor to built-in sensors, e.g., in the tinnitus domain,

we are experimenting with mobile EEG sensors. TinnituSense is

faced with several challenges of which we already solved principle

architecture, sensor management, and energy consumption.

Index Terms—mobile crowdsensing; tinnitus; automated mo-

bile sensing; mobile health;

I. INTRODUCTION

Tinnitus is a common disorder, which is associated with

the perception of a ringing sound or noise in the ears. The

causative factors of tinnitus are still unknown, however, it

usually is intertwined with an underlying condition in the

ear. Little is also known about which external conditions, e.g.,

weather, and how they influence a patient’s tinnitus experience.

Recently, smartphone apps and health appliances like heart

and step meters got popular to help patients to maintain and

mitigate their health problems. Smartphones are especially

interesting as they nowadays come with high computational

power, pose a good battery life time, and a set of built-in

sensors that can easily be managed by apps. Even if there is a

need for additional sensors, they can easily be connected (e.g.,

via BlueTooth). Further, smartphones come with an application

ecosystem that can be extended by new apps, which are

just software that can be programmed for a particular health

problem. Finally, users are typically familiar to the touch-

based interface, which can be used for customizations, but

also to allow arbitrary input (e.g., questionnaires).

There are apps collecting patient data by asking for the

patients’ conditions (participatory sensing), which are relevant

for therapists tracking these conditions during their patients’

daily routine. In case of tinnitus, patients can fill-in app-

presented questionnaires that can be analysed by researchers

and therapists in order to gain new insights into tinnitus.

Other apps can continuously track patients’ behaviour, like

steps gone or heart-beat rates during the day. However, to the

best of our knowledge, there is no app yet that continuously

and automatically tracks patients’ behaviour or other charac-

teristics around patients (e.g., pressure, light, loudness, etc.)

and correlates those to the patients’ own condition (e.g., their

annoyance or affection by tinnitus).

This paper presents TinnituSense, our novel concept and

proof-of-concept implementation of an app tracking weather

conditions and correlating them to the current patient condition

by filing in a questionnaire. Researcher and therapists can

use collected data for a per-patient analysis, but also for

broader analyses of how tinnitus is influenced by weather

conditions in general. Our concept is applicable to other

domains, but focussed on tinnitus as it was developed as part

of the European School for Interdisciplinary Tinnitus Research

(ESIT)1.

We organised this paper as follows: Section II reports on

related work. In Section III, we will give some background

and motivation for our work. While Section IV describes our

approach, its implementation, and our solution of challenges.

Then, in Section V, we present first results and their impact.

Finally, Section VI concludes our work and provides some

future considerations.

II. RELATED WORK

With respect to employing mobile crowdsensing in the

healthcare domain, various solutions have been proposed in

literature. For instance, in case of individual-centric ap-

plications, Györbíró et al. propose an activity and posture

detection mechanism [6]. A more detailed use-case–based

1funded by the EU Horizon 2020 program as Marie Skłodowska-Curie

grant, agreement number 722046

solution, specifically targeted towards activity detection of

patients suffering from Parkinson’s disease is reported in [2].

Additionally, the authors in [17] and [14] report image-based

and video-imaging–based solutions for diet and heart-rate

monitoring. An attempt to use social-media data to monitor

emotional health of patients is presented in [10].

Similarly, in case of environment-centric applications, [4]

presents an approach that enables use of specialised auxiliary

sensors to achieve participatory sensing for active monitoring

of air quality and pollution. Authors of [8] present a location-

and machine-learning–based solution towards observance of

disease and disaster outbreak. Both of these approaches rely

highly on participatory sensing coupled with minor automated

sensing, carried out by auxiliary or specialised sensors.

Within the scope of tinnitus research, a plethora of scientific

literature report different applications of mobile crowdsensing,

ranging from data collection to mitigating tinnitus symptoms

and supporting clinicians. For instance, the smartphone app

TrackYourTinnitus, based on participatory sensing, systemati-

cally records fluctuations of tinnitus symptoms from patients.

With these records a connection between tinnitus and daily

routine or activities [15] can be established. Unlike Track-

YourTinnitus, our TinnituSense approach supports automated

sensing of weather-related data in addition to feedback (in

form of THI surveys) from the patients.

NoiseTube [11] is another pertinent app that acquires geo-

localised noise levels using the embedded microphone of the

smartphone. This data is therefore examined by scientists to

identify relationship between behavioural and psychological

problems associated with noise pollution. Similarly, Sound

Meter uses the microphone of mobile devices to measure the

environment loudness and gives a reference sound to compare

the loudness with [13]. Unlike both apps using only micro-

phone and GPS to measure the environment, TinnituSense uses

several additional sensors, such as the photometer, barometer,

thermometer, hygrometer, accelerometer, and gyroscope to get

additional information about the environmental condition.

III. BACKGROUND AND MOTIVATION

A. Tinnitus

As mentioned, tinnitus is a disorder associated with the

perception of irritating sound or noise. In addition to general

health complications, tinnitus might be also responsible for

provoking other psychological disorders (e.g., stress, anxiety,

depression, or obsessive-compulsive disorder) and may affect

the common as well as social lifestyles. Furthermore, tinnitus

has also been described to may have correlations with migraine

and vertigo. The perception of tinnitus is driven by the filtering

work of subconscious areas in the brain-stem influenced by the

limbic system, and increases in episodes of interior or exterior

stress.

Changes in the atmospheric surrounding may have direct

or indirect effects on the annoyance caused by tinnitus. Some

scientific literature has reported an alleviation in the tinnitus

condition of patients with Meniere’s Disease, which can be

related to the surrounding atmospheric and environmental

conditions [19]. Among others, factors like a decrease in the

atmospheric pressure, a weather change (specifically rainy and

cold weather), intensity of light, the current sound environ-

ment, or (sudden) change in altitude are some of the most

pertinent ones. Importantly, weather is no causative factor for

tinnitus, but may influence the perception of the sounds. Very

often patients report that they hear intermittent secondary tones

to their common tinnitus sound or describe their common tone

changing from compensated to annoying expressed by terms

like hammering, beating, fizzling.

Similarly, air pressure is considered to be responsible for

episodes of migraine in many patients [9]. In rare cases,

patients describe a change in tinnitus perception corresponding

to their migraine in major air pressure changes. On the other

hand, the orientation, movement speed, light intensity, and

direction of movement of the patient are some of the less

common, yet still significant set of factors that may also induce

a spike in tinnitus symptoms.

Tinnitus patients often report that the perception of their

tinnitus varies. The tinnitus loudness, as well as the tinnitus-

related distress, can change from one moment to the other.

This moment-to-moment variability can be measured with

Ecological Momentary Assessment methods [18]. Even though

some influencing factors for this variability has been identified

[16], a large amount of the variability cannot be explained.

With this project, we also want to investigate the influence of

weather conditions on the individual perception of tinnitus.

B. TinnituSense Approach

We designed TinnituSense as a concept for mobile sensing

and crowdsensing to enable advanced tinnitus research. Fig. 1

shows the actors in the TinnituSense use cases: patients,

researchers, therapists, and a backend storing collected data.

In [12], we presented a reference architecture that covers all

aspects of a TinnituSense-like application, e.g., mobile sensor

management, local and backend services, data management,

analytics, etc. As mobile sensing can be applied to monitor

the aforementioned atmospheric factors (most, if not all) that

provoke or influence tinnitus, this paper presents a proof-of-

concept app for this purpose. Our contribution compared to

previous work is the correlation of weather conditions to the

condition of the patient. This is done by not only automatically

collecting sensor data, but to also connecting them to the

patient’s current experience by using the Tinnitus Handicap

Inventory (THI)2, a standardized questionnaire.

In principle, TinnituSense could be extended to further

aspects: (i) The app could identify (drastic) changes in the

circumstances (motion, environment, and position) of a patient

who is prompted to fill-in a questionnaire about the severity

of their current situation. These circumstances could be pre-

configured by researchers or therapists. (ii) In detected cir-

cumstances the app could, configured by therapists, help to

mitigate problems, e.g., by acoustic therapy (calming sounds,

music therapy, listening pink or brown noise), intake of

2http://www.tinnituslab.com/Questionnaire_THI_ENGLISH.pdf

Fig. 1: TinnituSense Principle Actors

supplements, or personalised sound frequency treatments. (iii)

Feedback to the patients about correlations may keep the

patient aware of specific triggers and what may cause their

tinnitus symptoms.

However, applying smart mobile crowdsensing in real-

world scenarios brings forward a couple of technical chal-

lenges. Even though modern smart phones provide sophisti-

cated hardware, their manufacturers intended to make these

smartphones general purpose appliances. Herein, for sensing

in the healthcare domain, it is pertinent to acquire accurate

data, and since the sensors embedded in smartphones are not

dedicated, the problem of inaccurate sensor data arises [7],

in addition to the reliance on the sensed data [3]. In order to

achieve more accurate sensed data, continuous sampling of the

sensor data is required, which results in battery consumption

problems [21]. Moreover, as the smart phones are resource

constrained, employing smart mobile crowdsensing introduces

scarcity of hardware for general user experience [1]. And

in order to gain full potential of the mobile crowdsensing

application, it is pertinent that the sensed data is transmitted to

some cloud or backend, where it is further processed. This, in

turn, introduces the data transmission challenges [5], as well

as its associated energy consumption problem [20].

IV. IMPLEMENTATION

The current implementation of the TinnituSense App has

been carried out for Android platform using Google’s dedi-

cated Integrated Development Environment (IDE)—Android

Studio3. The current version of the app uses a combination of

Android’s standard libraries as well as OnsenUI4to provide

rich user experience. Currently, TinnituSense supports the

following sensors:

•Photometer (light sensor): Measures the luminance in the

unit Lux (lux)

•Barometer (air pressure sensor): Measures the ambient

air pressure in the unit Hectopascal (hPa)

•Thermometer (temperature sensor): Measures the ambient

air temperature in the unit Degree Celsius (°C)

•Hygrometer (humidity sensor): Measures the ambient

relative humidity in percent (%)

3https://developer.android.com/studio/

4https://onsen.io/

Fig. 2: A. Home Screen - B. Sensors Page - C. Sensor Detail

•Accelerometer (acceleration sensor): Measures the veloc-

ity along the x-, y- and z-axis (including gravity) in the

unit Meters per Second Squared (m

s2)

•Gyroscope (angular rate sensor): Measures the rate of

rotation around the x-, y- and z-axis in the unit Radiant

per Second (rad

On first time invoking the app, it initially probes the embedded

sensors and identifies the supported sensors by the device. In

case, if any of above mentioned sensors is not supported, the

app displays the ’Not Supported’ message to the user. These

settings are then stored locally to avoid future probes. Further-

more, the user gets a ’Home’ Screen (cf. in Fig. 2(A)) with

four navigational buttons. Each button has textual descriptions

as well as pictorial representations of the underlying function.

Additionally, an easy-access navigational sub-menu button on

the upper right corner is accessible throughout every screen in

the app. The sub-menu enables quick access to aforementioned

four navigational buttons as well as the home screen.

The Survey button (cf. Fig. 2(A)) invokes the Tinnitus

Handicap Inventory (THI) questionnaire within the app. If a

user completes and submits answers to the survey, the app

scores the answers based on the THI scale. This score, answers

to the survey questionnaire, timestamp of survey response are

then stored primarily on the device, and once a data connection

has been established based on user’s preference (WiFi or

Mobile), this data is then transmitted to a backend. In addition

to wilfully filling out the survey at any time, the user is also

periodically prompted via app notifications.

The Settings enable app customizations, like enabling day

or night mode as a theme. Enabling or disabling app no-

tifications, changing the frequency of THI survey prompts

(e.g., weekly, monthly etc.), and data transmission options,

for instance via WiFi connection or mobile data connection.

Since TinnituSense promotes crowdsensing, it is pertinent

that individual profiles are generated to uniquely identify

patients. The Profile section of the app thus acquires user-

specific data. This data is sub-divided into two major parts,

primarily, the user provides basic (so called ’non-personal’)

information about themselves like age and gender. We ensure

that we do not ask any personal questions like name, and date

of birth. Secondarily, the app asks the user about tinnitus-

related questions, for instance, first encounter with tinnitus,

type and severity of tinnitus, therapy, and medication related

questions. In addition to having a user-specific profile, this

information is also used to generate a unique profile identifier.

As sensors are important participants of the app, information

specific to their current activity is hence deemed critical.

This is therefore, tapping the Sensors button (cf. Fig. 2 (A)),

will lead the user to list of sensors supported by the device,

screenshot of this is depicted in Fig. 2 (B). Herein, each

supported device sensor is shown using a text and a pictorial

representation (icon), furthermore, the user can also see the

current sensor value, as well as has the option to turn a device

sensor ON or OFF. Since, we do not automatically manage

the activation and deactivation of the sensors within the app,

we leave it to user’s judgement to enable and disable a certain

sensor based on its battery consumption and significance.

We also identified that the user would be interested in

visualising the sensor values to gain an insight into his current

surrounding environment. In terms of usability, provision of an

observable action by the sensors would gain user’s trust in the

application. Therefore, tapping on any sensor in Fig. 2 (B),

will lead the user to sensor-specific page. An example of such

a page for ’Light Sensor’ is shown in Fig. 2 (C). The sensor

specific page shows the live sensor value in addition to a line

chart with the history of the most recent measurements. The

chart is updated with every new measurement and the ranges

on the x- and y-axes are adjusted dynamically. We believe

that the line chart would be sufficient for a user to properly

interpret the sensor values, enabling him to identify peak and

low values and compare with current tinnitus sensation. The

user also has the option to activate or deactivate a sensor within

the sensor specific page. And a timestamp showing the time

the sensor was activated is also shown.

As mentioned before, the activation and deactivation of a

sensor is dependent on user’s judgement. We employ this

scheme to provide users with sufficient freedom to control

critical aspects of the app. However, we ensure to support

user’s judgement by provision of necessary values regarding

the battery consumption of individual sensor. The different

variations of available information regarding the battery con-

sumption of individual sensor are shown in Fig. 3. Where

Fig. 3 (A) provides information on current battery usage and

a bar-chart showing the battery consumption of the sensor

for past 5 activation periods, both values are given in mAh.

Every new bar in the chart is rendered at the moment the

user turns the corresponding sensor off. Fig. 3 (B), shows a

pie-chart comparing the current battery consumption of the

sensor with overall capacity of the device battery, herein, the

red line signifies the power consumed by ’Light Sensor’ since

its last activation, while the grey colour signifies the battery

capacity (dark grey for available battery, and light grey for

used battery). And in Fig. 3 (C), we show a pie-chart depicting

the current battery usage of the sensor in relation to other

active sensors of the device as well as the remaining and

used battery of the device in mAh. The legend at the bottom

describes the colour codes associated with their respective

device sensors. Both of these pie-charts are updated at the

Fig. 3: Battery Usage :- A. Last 5 activations B. In relation to device battery

C. In relation to other sensors and device battery

interval of 30 seconds.

We argue that the battery consumptions of individual sen-

sors, in relation to the device battery, and other sensors will

provide the user with necessary information to support the

decision of turning a sensor ON or OFF. For instance, from

Fig. 3 (C), the user can easily identify that the gyroscope is

consuming large amount of battery, and if the user deems that

the gyroscope is not significant in relation to his tinnitus, he

can decide to turn the gyroscope sensor OFF. Similarly, this

will enable the user to experiment with multiple sensors while

conserving the battery to identify sensors which are critical

in terms of his tinnitus sensation. For example, for a user

whose tinnitus symptoms worsen due to changes in altitude,

the barometer sensor will play a significant role.

V. RESULTS AND DISCUSSIONS

To check the applicability, usability, and practicality of the

proposed approach and the app, we performed both quantita-

tive and qualitative tests. For quantitative tests, we used two

standard smartphones in default configuration, Google Nexus

4with Android version 6.0 and Samsung Galaxy with Android

version 4.4.

As, it is possible for a user to turn ON and OFF the device

sensors using the app, an indication of individual sensor’s

battery (power) consumption can be helpful in managing the

over-all battery of the device. To better understand and to

get an indication of how much battery capacity is consumed

by the individual sensors of the two devices, four different

tests were carried out over a period of 12 hours. An average

battery consumption of individual sensor per device is shown

in Table I. It is evident that the Barometer of both devices

uses the least energy, while the Gyroscope consumes the

most. In all test runs, the battery consumption is determined

by activation time (in ms) multiplied by strength of electric

current of the corresponding sensor (in mAh).

Average battery usage in mAh per 12h

Sensor Google Nexus 4Samsung Galaxy

1. Photometer 2.11 8.95

2. Barometer 0.07 0.26

3. Thermometer unsupported 7.96

4. Hygrometer unsupported unsupported

5. Accelerometer 6.1 1.68

6. Gyroscope 43.28 73.51

TABLE I: Device - Sensor: Battery consumption

Fig. 4: Patient Demographics and Tinnitus Durations

Herein, since it could be argued that the sensor’s battery

consumption is subjective to the amount of its usage (for

instance, the data it collects and forwards to the app), therefore,

we ensured that the battery consumption of each sensor is

computed while the sensor is completely activated in an ideal

environment, and also probing it periodically for data. For

example, for light sensor (photometer), Android system only

activates and shares the sensor data when it detects changes in

light intensity. Therefore, for testing purposes, we ensured that

sensor was exposed to different light intensities and in addition

to data being pushed from the light sensor, we implemented a

pull mechanism that periodically requested data from sensor.

During these tests, another keen observation was that the two

devices generated varying measurements for some sensors in

same circumstances. We believe that this could be resulting

from the sensors being differently calibrated on both devices.

Further, we wanted to highlight the impact and need of a

mobile crowdsensing app, like TinnituSense. We were inter-

ested in the opinion of the patients in terms of using such

an app, and wanted to identify the reality between weather-

related changes and their impact on tinnitus symptoms from

the perspective of a patient. Finally, we liked to see the

significance of medication or therapy-related solutions. There-

fore, we conducted qualitative tests in form of 10 personal

interviews with patients suffering from tinnitus at ENT Clinic

of the German Army Hospital of Ulm.

The demographics of these patients, in terms of age, gender,

and the duration of tinnitus are given in chart depicted in

Fig. 4. The x-axis represents the duration or how long a patient

has been suffering from tinnitus, while the y-axis gives the

number of patients. Gender is separated by different colour

codes, while the age (in years) is shown within individual

blocks. For instance, there are 2 patients, 1 male aged 27 and

1 female aged 31, who have been suffering from tinnitus for 6–

12 months. Within these diverse set of interviewees, 8 of them

reported their tinnitus being caused by stress, while 3 also

mentioned sound exposure. There was 1 infection-related case,

and 1 listed their tinnitus caused by unknown reasons. On

asking about their tinnitus severity on the scale of light, mild,

moderate, severe, and catastrophic, 2 patients reported mild,

5 moderate, and 3 patients reported severe tinnitus burden.

80% of the patients reported accompanying dizziness, while

50% mentioned hearing loss caused by tinnitus.

None of the patients had ever taken a THI questionnaire.

Among those patients, 50% reported that their tinnitus symp-

toms worsen due to changes in weather (2 reported sum-

mer being critical, 2 reported winter, and 1 patient reported

drastic changes to be elevating factor in tinnitus symptoms).

Furthermore, on asking about their general (job, shopping,

cooking) and social (party, gatherings, dinners) activities, 50%

of the patient responded that they reduce their general activity,

while 70% reported avoiding social activities. Out of 10,

a total of 6 people reported use of medication to control

tinnitus, among these, only 5 patients described medicines

to be useful. Similarly, on asking if therapeutic solutions

helped, all patients responded positive. Additionally, we asked

patients if they used smartphones or any specific apps to

help control their tinnitus, every patient responded positive

to having a smartphone, while all of them responded negative

to use of any application. Based on this, we asked if they

would be interested in using such an app (while explaining the

TinnituSense app), 9 out of 10 patients agreed. And all these 9

out of 10 patients agreed to sharing their tinnitus related data

for research purposes.

VI. FUTURE WORK AND CONCLUSION

The current version of the TinnituSense app utilizes sensors

which are embedded in most commonplace smartphones, and

attempts at acquiring environmental data and monitors patient

activity using the aforementioned sensors. This can help in

understanding correlations between environmental or weather

related changes with tinnitus symptoms. However, within the

context of tinnitus research, there is still room to implement

auxiliary sensors. For instance, mobile Electroencephalogra-

phy (EEG) systems coupled to smartphones to acquire elec-

trical activity of the brain, and enabling provision of real-

time neurofeedback to the patients suffering from tinnitus.

Additionally, the data acquired from mobile EEG systems can

also be helpful for researchers and psychologists to better un-

derstand tinnitus and its associated brain activity. To establish a

correlation between weather and tinnitus, the current version of

the app implements the THI questionnaire. Herein, we believe

that the app could benefit from adding tinnitus loudness and

tinnitus distress related questions from TrackYourTinnitus app,

as well as comparing the data from both apps. Currently, the

app generated data (sensors, questionnaires, profile, settings,

etc.) are stored locally on the device, and are transmitted to the

backend once a data connection is established. However, we

are presently working on a sophisticated Sensor Data Storage

and Transmission technique by means of employing 1) Ad-

hoc Storage and 2) Distributed Processing. Where, Ad-hoc

Storage will ensure storage of acquired sensor data on the

device in compressed and optimal format in absence of data

connection. While, Distributed Processing will find a balanced

distribution of raw sensor data to be processed by the device

and the backend. This, in turn, will enable better device battery

life conservation, reduction of data transmission overheads and

save processing resources of the device.

Furthermore, even though the current implementation of

TinnituSense attempts at conserving energy, there is still

significant room for improvements. For instance, as of now,

the app relies on Android’s internal power management and

depends on user’s own will to manage individual sensor’s

battery consumption by turning it on/off manually. Herein,

an automated sensor management technique, which identifies

non-significant use of sensor and turns it off, as well as

activation of high battery consuming sensors based on a

criteria can further benefit the battery conservation. Moreover,

to improve manual intervention of user to conserve battery

by turning a sensor off (cf. Table I showing high battery

consumption by the gyroscope), a notification scheme that

reminds the user that a sensor with high battery consumption

has been running for a longer period of time could be used,

so that the user can decide whether or not it would be better

to turn off this sensor to save energy. As we highlighted

previously, that the embedded sensors of smartphones are gen-

eral purpose and therefore prone to data accuracy problems,

the app could benefit from employing techniques that could

drastically improve the data accuracy issues. For instance,

achieving sensor fusion of temperature sensor by comparing it

with third-party mobile apps (for example, Google Weather),

could help in identifying error delta and thus re-calibrating the

sensor can improve accuracy.

In this paper, we highlight the significance of mobile

crowdsensing in the healthcare domain, specifically in tinnitus

research. We reported the lack of automated sensing apps

for tinnitus research and identify the technological gap that

limits scientific studies to correlate environmental and weather

changes to tinnitus symptoms. Therefore, we presented an ap-

proach towards automated sensing of environmental conditions

and patient monitoring, backed by the use-case of tinnitus

research. The approach is sustained by implementation of a

smartphone app, a user-interface, and implementation of a use-

case specific questionnaire. We believe that the technological

steps undertaken in this paper will help in carrying out large

scale scientific studies to quantify data correlating tinnitus

symptoms with weather related and environmental changes.

Results of a small survey on patients gives hope that patients

are interested in using such an app. Thus, we believe that

enabling the technology is one of the significant contributions

of this paper. Furthermore, we argue that a sophisticated

mobile crowdsensing app designed for monitoring patient

surroundings and profiling patients for personalised therapy

may foster controlling and mitigating of tinnitus symptoms,

as well as promote community or participatory sensing.

REFERENCES

[1] Saeid Abolfazli, Zohreh Sanaei, and Abdullah Gani. Mobile cloud

computing: A review on smartphone augmentation approaches. arXiv

preprint arXiv:1205.0451, 2012.

[2] Mark V Albert, Santiago Toledo, Mark Shapiro, and Konrad Koerding.

Using mobile phones for activity recognition in parkinson’s patients.

Frontiers in Neurology, 3:158, 2012.

[3] Jeffrey R Blum, Daniel G Greencorn, and Jeremy R Cooperstock.

Smartphone sensor reliability for augmented reality applications. In

Int. Conf. on Mobile and Ubiq. Sys.: Comp., Netw., and Services, pages

127–138. Springer, 2012.

[4] Prabal Dutta, Paul M Aoki, Neil Kumar, Alan Mainwaring, Chris Myers,

Wesley Willett, and Allison Woodruff. Common sense: participatory

urban sensing using a network of handheld air quality monitors. In

Proc. of the 7th ACM Conf. on Emb. Netw. Sensor Sys. (SenSys), pages

349–350. ACM, 2009.

[5] Raghu K Ganti, Fan Ye, and Hui Lei. Mobile crowdsensing: current

state and future challenges. IEEE Comm. Mag., 49(11), 2011.

[6] Norbert Györbíró, Ákos Fábián, and Gergely Hományi. An activity

recognition system for mobile phones. Mobile Netw. and App., 14(1):82–

91, 2009.

[7] Mohammad Ashfak Habib, Mas S Mohktar, Shahrul Bahyah Ka-

maruzzaman, Kheng Seang Lim, Tan Maw Pin, and Fatimah Ibrahim.

Smartphone-based solutions for fall detection and prevention: challenges

and open issues. Sensors, 14(4):7181–7208, 2014.

[8] Peter Haddawy, Lutz Frommberger, Tomi Kauppinen, Giorgio De Felice,

Prae Charkratpahu, Sirawaratt Saengpao, and Phanumas Kanchanakit-

sakul. Situation awareness in crowdsensing for disease surveillance in

crisis situations. In Proc. of the 7th Int. Conf. on Inform. and Comm.

Techn. and Dev. (ICOICT), page 38. ACM, 2015.

[9] Kazuhito Kimoto, Saiko Aiba, Ryotaro Takashima, Keisuke Suzuki,

Hidehiro Takekawa, Yuka Watanabe, Muneto Tatsumoto, and Koichi

Hirata. Influence of barometric pressure in patients with migraine

headache. Internal Medicine, 50(18):1923–1928, 2011.

[10] Mark E Larsen, Tjeerd W Boonstra, Philip J Batterham, Bridianne

O’Dea, Cecile Paris, and Helen Christensen. We feel: mapping emotion

on twitter. IEEE J. Biomed. and Health Inf., 19(4):1246–1252, 2015.

[11] Nicolas Maisonneuve, Matthias Stevens, Maria E Niessen, and Luc

Steels. Noisetube: Measuring and mapping noise pollution with mobile

phones. In Inform. Techn. in Environm. Eng., pages 215–228. Springer,

2009.

[12] Muntazir Mehdi, Guido Mühlmeier, Kushal Agrawal, Rüdiger Pryss,

Manfred Reichert, and Franz J.Hauck. Referenceable mobile crowd-

sensing architecture: A healthcare use case. In Proc. of the 1st Int.

Worksh. on Serv. for Mobile Data Coll. (MoDaC), volume 134, pages

445–451. Procedia Computer Science, 2018.

[13] Enda Murphy and Eoin A King. Testing the accuracy of smartphones

and sound level meter applications for measuring environmental noise.

Applied Acoustics, 106:16–22, 2016.

[14] Ming-Zher Poh, Daniel J McDuff, and Rosalind W Picard. Non-contact,

automated cardiac pulse measurements using video imaging and blind

source separation. Optics Express, 18(10):10762–10774, 2010.

[15] Thomas Probst, Rüdiger Pryss, Berthold Langguth, and Winfried Schlee.

Emotional states as mediators between tinnitus loudness and tinnitus

distress in daily life: Results from the “TrackYourTinnitus” application.

Scientific reports, 6:20382, 2016.

[16] Thomas Probst, Rüdiger C Pryss, Berthold Langguth, Josef P

Rauschecker, Johannes Schobel, Manfred Reichert, Myra Spiliopoulou,

Winfried Schlee, and Johannes Zimmermann. Does tinnitus depend

on time-of-day? an ecological momentary assessment study with the

trackyourtinnitus application. Frontiers in aging neuroscience, 9:253,

2017.

[17] Sasank Reddy, Andrew Parker, Josh Hyman, Jeff Burke, Deborah Estrin,

and Mark Hansen. Image browsing, processing, and clustering for

participatory sensing: lessons from a dietsense prototype. In Proc. of

the 4th Worksh. on Emb. Netw. Sensors (EmNetS), pages 13–17. ACM,

2007.

[18] Winfried Schlee, Rüdiger C Pryss, Thomas Probst, Johannes Schobel,

Alexander Bachmeier, Manfred Reichert, and Berthold Langguth. Mea-

suring the moment-to-moment variability of tinnitus: the trackyourtin-

nitus smart phone app. Frontiers in aging neuroscience, 8:294, 2016.

[19] Wiebke Schmidt, Natalie Sarran, Christophe an Ronan, George Barrett,

David J Whinney, Lora E Fleming, Nicholas J Osborne, and Jessica

Tyrrell. The weather and meniere’s disease: a longitudinal analysis in

the uk. Otology & Neurotology, 38(2):225, 2017.

[20] Haoyi Xiong, Daqing Zhang, Leye Wang, and Hakima Chaouchi. Emc

3: Energy-efficient data transfer in mobile crowdsensing under full

coverage constraint. IEEE Trans. on Mobile Comp., 14(7):1355–1368,

2015.

[21] Zhenyun Zhuang, Kyu-Han Kim, and Jatinder Pal Singh. Improving

energy efficiency of location sensing on smartphones. In Proc. of the

8th Int. Conf. on Mobile Sys., App., and Services (MobiSys), pages 315–

330. ACM, 2010.