Sensors are Power Hungry: An Investigation of Smartphone Sensors Impact on Battery Power from Lifelogging Perspective

Article (PDF Available) · December 2016with 621 Reads
Cite this publication
Abstract
Smartphones are ubiquitous devices with millions of units sold around the globe every year. To meet the growing performance needs of innovative applications, smartphones industry has mainly exploited the rapid technological developments in computing, storage and communication with lesser regard to the advancements in battery technology. Sensors integration has turned smartphones into powerful sensing and computing devices with unlimited opportunities for devising novel methods of sensing. Smartphones sensing has provided novel computing methods including human computer interaction and context awareness which is of immense importance in different fields including lifelogging. However, continuous sensing can consume significant amount of energy and inefficient use of sensors is a major source of smartphones energy consumption. In this paper, we have articulated the importance of smartphone sensors in users' daily life activities and their usage effect on a smartphone's battery life. For this purpose, an Android app Energy Monitoring System for Smartphones Sensors (EM3S) is designed. EM3S is experimented in several real world scenarios for estimating smartphones sensors energy consumptions information producing a huge amount of useful data for potential exploitation in fields like lifelogging. It was found that different sensors consume different amount of energy during different daily life activities which can affect smartphones performance and is a major hurdle for sensors based applications. This study will help device researchers, manufacturers and developers in exploring optimal sensors energy consumption methods while developing pragmatic sensors based systems.
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 8 ISSN 1999-4974
Sensors are Power Hungry: An Investigation of Smartphone Sensors
Impact on Battery Power from Lifelogging Perspective
Inayat Khan, Shah Khusro, Shaukat Ali, and Jamil Ahmad
Abstract Smartphones are ubiquitous devices with
millions of units are sold around the globe every year. To meet
the growing performance needs of innovative applications,
smartphones industry has shown tremendous developments in
computing, storage capacity, communication, and battery
power technologies. The integration of sensors has turned
smartphones into powerful sensing methods with unlimited
opportunities for devising novel applications for solving real-
world problems. This has given rise to a new area of research
called smartphone sensing which have potential applications in
different domains including lifelogging. However, continuous
and inefficient usage of sensors in lifelogging can consume
significant amount of battery power and can drain out fully
charged battery within a few hours. In this paper, we have
presented the importance of smartphone sensors in monitoring
users' daily life activities and their usage effects on smartphone
battery lifetime. For this purpose, an Android app Energy
Monitoring System for Smartphones Sensors (EM3S) is
developed. EM3S is experimented in several real world
scenarios for estimating smartphones sensors battery power
consumptions information. It is found that smartphone sensors
consume varying amount of battery power during different
daily life activities. However, collectively, they can affect
smartphone performance and is a major hurdle for smartphone
sensors-based applications. This study is aimed to help
researchers, manufacturers and developers in exploring
optimal sensors battery power consumption methods while
developing pragmatic smartphone sensors-based applications.
Index Terms Context-Awareness, Lifelogging, Power
Consumption, Sensors, Smartphone.
I. INTRODUCTION
The advancements in science and technology have
empowered semiconductor technology to manufacture low-
cost, high-power, and multi-functional mechanical devices
called chips.
In general, following Moore's law, the number of
transistors in a unit area doubles after each eighteen months
and smartphones goes one step forward by fabricating more
and more functionalities in a single chip to compensate
budget [1]. Recent developments in the sensors’ issues such
as size, processing requirements, and cost effective
production have enabled sensors integrations in products and
appliances [2].
Smartphones are modern high-end mobile phones
combining the features of pocket sized communication
devices with PC like capabilities [3]. Smartphones are
powered with powerful hardware and sophisticated operating
systems that enable them to execute sophisticated even
scientific applications covering a wide variety of domains
and store as well as process a large volume of data [4]. It was
formalized that extending sensory technology to
smartphones could substantially increase their capabilities
and functionalities. Smartphones sensing capability includes
a rich set of specialized sensors (i.e., GPS, accelerometer,
proximity, gyroscope, magnetometer, microphone, Wi-Fi,
and ambient light etc.). [5]. Incorporation of sensors in
smartphones has changed their role from traditional
communication devices into life-centric sensors [6]. Sensing
capabilities enables smartphones to unobtrusively monitor
and accumulate a broad range of dynamic information about
people's physical activities [5, 7], contexts-awareness and
environmental conditions [8, 9] etc., in real time.
Increasing incorporation of sensors in smartphones
fosters the proliferation of various sensors-based
applications. The smartphones ambient sensing power can be
used as a primary tool for providing context information to a
new class of smartphones cooperative services [1]. Sensors-
based applications can sense a user’s environment and
provides effective context-aware services [9] such as Google
Maps can use Global Positioning System (GPS) sensor to
provide location-aware services to navigate hikers in a rural
area, and accelerometer sensor can aid functionalities to
games and photography etc. Lifelogging is a special bread of
context-aware applications that emphasizes on the creation
of surrogate memory (digital archive) of a person's lifetime
experiences by continuously and unobtrusively capturing
and storing of contextual information about his daily life
activities. Lifelogging systems urge on the use of sensory
technology for ambient sensing of contextual and
environmental information about users, and using of the
captured information as cues to augment their episodic
memories. Ambient sensing of contextual information for
lifelogging can bring applications' capabilities to new level
of sophistication such as providing memory aids to the
peoples suffering with cognitive memory impairments (e.g.,
Alzheimer, and Amnesia etc.) etc. The sensory capabilities
make smartphone as a suitable lifelogging device. However,
accurate context identification needs accurate measurement
of context features including motion, background condition,
and location etc., which are resources intensive tasks.
The increase in smartphones sensing capabilities has
raised power need issue to a level which could not be met by
the current smartphones limited power source. In
smartphones, battery size and capacity is severely restricted
due to size and weight constraints of the devices [3]. A
smartphone featuring with a conventional cellular radio
antenna, collection of sensors and services, touch screen, and
many others requires greater power source because each one
is taking toll on the limited battery resource [10]. Empirically
Inayat Khan, Shah Khusro and Shaukat Ali are with Department of Computer Science,
University of Peshawar, KP, Pakistan and Jamil Ahmad is with the Department of
Computer Science, Research Center for Modeling and Simulation (RCMS), National
University of Sciences and Technology, Pakistan.
Email:inayat_khan@upesh.edu.pk,khusro@upesh.edu.pk,shoonikhan@upesh.edu.pk,
dr.ahmad.jamil@gmail.com.
Manuscript received Feb 03, 2016; revised on Sep 15, 2016 and Nov 25, 2016;
accepted on Dec 12, 2016.
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 9 ISSN 1999-4974
applications using sensor can be the root cause of power
wastage by failing in determining the effective use of sensors
and their data [11]. Therefore, sensors are needed to be used
cost-effectively otherwise would result in complete battery
drain quickly [11]. Smartphone limited battery power can
foster big hurdles and restrictions for smartphone-based
lifelogging applications which require huge power due to
using sensors. Researchers have investigated power
consumption optimization at different levels (i.e., hardware
and software etc.) and defined power management strategies
either by immediately shutting down of unnecessary sensors
or by carefully alignment of sensors duty cycles [1].
However, suggesting an effective strategy requires prior
insight knowledge of different smartphone sensors power
consumption rates. Such precise knowledge would also
enable lifelogging applications developers to employ sensors
on where and how philosophy in order to save power while
producing qualitative lifelogging applications without
jeopardizing the underlying platforms.
In this paper, we have presented the importance of
smartphone sensors in monitoring users' daily lives activities
and their usage effects on smartphone battery lifetime. To
help in our investigation, an Android app namely Energy
Monitoring System for Smartphones Sensors (EM3S) is
developed to effectively monitor, record, and analyze the
power consumption rates of the various smartphone sensors.
For estimating energy consumption rates of each sensor
explicitly and in conjunction with other sensors, an extensive
test criteria has been defined which consists of different real
world scenarios. All of the tests in each of the scenarios are
carried out using EM3S on QMobile A12 smartphone.
Results obtained have revealed that smartphone sensors can
consume excessive amount of battery power during tasks
completion; therefore, can be the major source effecting
smartphone battery lifetime. In addition, sensors have
variable power consumption rates where some sensors
consume less and the others consume very much battery
power.
II. ANDROID POWER MANAGEMENT SYSTEM
Android implements a mechanism to prolong battery
life. When an Android device is left idle, it will first dim,
then turn off the screen, and, finally turn off the CPU.
Android provides a dedicated power management API in the
Applications Framework layer which can be accessed by the
running applications and services using Wake Locks Power
Manager system service to control the power state of the host
device. Android provides four Wake Locks types where
each Wake Lock type determines CPU, screen lightness, and
keyboard lightness as shown in Table I.
PARTIAL_WAKE_LOCK is used by the Android services
which run in the background and have no user interface for
users’ interactions. CPU will be shut down if no Wake Lock
is active. An active Wake Lock, depending on its type,
thwarts device from suffering full system suspend state (.i.e.,
WAKE_LOCK_SUSPEND) or low-power state (.i.e.,
WAKE_LOCK_IDLE) [12].
When an application is launched, it initiates a new Wake
Lock by requesting CPU for Power Manager API in the
Application Framework which creates a Wake Lock and
transfers the lock request to the Power Management service
contained in the Linux kernel. The Power Manager also
response back to the application about Wake Lock creation
and signifies resources consumption depending on the Wake
Lock type created. Fig. 1 depicts the Linux modified internal
power management framework for Android devices with
limited battery power.
III. SMARTPHONES AND SENSORS BACKGROUND
The growing adoptability of smartphones by people and
recent technological developments has paved the way of a
new sensing paradigm by embedding a number of
specialized sensors in smartphones. Today’s smartphones
have several high valued embedded sensors that are having
rich sensing capabilities. In addition, smartphone can also
communicate with external sensors using wireless
networking protocols (e.g., Bluetooth etc.). To exploit the
rich sensing and technological capabilities of smartphone,
research community and industry have envisioned several
high valued applications for solving real world problems in
different domains such as health monitoring [14-16],
physical activities recognition and fall detection [5, 7, 17,
18], pollution monitoring [19], traffic monitoring and
automatic accident detection [8, 20], social networking [21]
etc. Using smartphone as sensor has several practical
advantages over traditional wireless sensor networks [19,
22].
Smartphones are always accompanied by users;
therefore, solves the problems of power management,
and network formation and maintenance.
Table I. Android Wake_Lock Options [13].
Wake Lock
CPU
State
Screen
Lightening
Keyboard
Lightening
FULL_WAKE_LOCK
Running
Full Bright
Backlight
Illuminated
SCREEN_BRIGHT_WAK
E_LOCK
Running
Full Bright
Backlight Off
SCREEN_DIM_WAKE_L
OCK
Running
Dim Light
Backlight Off
PARTIAL_WAKE_LOCK
Running
Off
Backlight Off
Fig. 1 Android power management architecture.
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 10 ISSN 1999-4974
Nodes in a wireless sensor network have relatively high
prices that increases the overall cost of a network
implementation. Using smartphone as sensors can have
high economy of scale as manufactured in large
quantities, and already owned by the users. All together,
could help in surpassing the overall cost in millions.
Smartphone can provide coverage to geographical areas
where static sensors are hard to deploy.
Smartphone can provide coverage where it is needed the
most and provide a close intact to the measuring
phenomenon to get accurate observations.
Human users assistance to smartphone can be used to
improve applications' functionalities such as camera can
be pointed appropriately by a human user to a target
object to be sensed.
A. Smartphones Sensors Classification
Sensors in smartphone can be categorized into physical
sensors and virtual sensors [23]. Physical sensors are
hardware-based sensors that are fabricated directly into
smartphone and derive their data directly by measuring a
particular environmental/contextual feature. For example,
accelerometer, gyroscope, and proximity etc. falls into the
category of physical sensors. Virtual sensors (also called
synthetic sensors or logical sensors) are software-based
sensors that are deriving their data by employing one or more
hardware-based sensors. For example, in Android platform
linear acceleration and gravity sensors are virtual sensors.
The number and types of sensors in smartphones varies
depending on the underlying smartphone platform and
usability. Understanding the potentialities of sensors and
increasing miniaturizations in technologies will enable the
integration of more advanced sensors in the future
smartphones [23]. Smartphone physical sensors can be
divided into two categories: general purpose sensors and
network interface sensors.
1) General Purpose Sensors
General purpose sensors either measures physical
properties related to the internal conditions or obtain
information about outside environmental/contextual
features. Each of the general purpose sensors captures
information about a particular topic which could be read and
analyzed by applications for effective decision making.
Some of the general purpose sensors available in modern
smartphones includes [23]: (1) proximity sensor detects any
nearby object in the electromagnetic field without any
physical contact, (2) accelerometer sensor can measure the
acceleration of a smartphone in 3-axis: X, Y, and Z to detect
orientation of the phone, (3) ambient light sensor can
measure light of the surrounds to optimize screen visibility
accordingly, (4) digital compass sensor recognizes the North
for identifying users directions, (5) gyroscope can measure
the position and orientation of a phone in 3-axis: yaw, pitch,
and roll, (6) Global Positioning System (GPS) sensor
receives geo-spatial information from GPS satellites and
calculate a user's location, (7) CMOS camera sensor uses
MOS (Metal Oxide Semiconductor) transistors to convert an
optical image into electrical signals, (8) microphone sensor
detects air pressure as vibration and creates an electrical
signal proportional to the vibration, and (9) temperature
sensor gives information about the ambient temperature
using solid state principles.
2) Network Interface Sensors
Network interface sensors are embedded sensors which
locate an external signal in the radio range, establish a
connection, and receive transmitted signals. The information
received by network interface sensors can be read by
applications for further usage. Each of the communication
sensors uses wireless networking technologies and protocols
for connecting with the remote objects (i.e., communication
devices or sensors etc.) using a particular frequency range of
electromagnetic spectrum at a specific data rate. Some of the
network interface sensors available in smartphone includes:
(1) Bluetooth sensor (IEEE 802.15.1) is a short range lower-
power broadcast communication sensor for connecting
personal consumer gadgets, peripherals, and sensors etc.
available in a proximity with a data rate less than 1Mbps, (2)
Wi-Fi sensor (IEEE 802.11) enables connectivity between
smartphone and a nearby (i.e., typically within 50 to 150
meters) Wi-Fi hot spot to provide high performance and
bandwidth of Wireless Local Area Network (WLAN) such
as Ethernet etc., and (3) Global System for Mobile
Communication (GSM) sensor enables connectivity and
maintenance with nearby BTS which in turn will be
connected to MSC.
B. Role of Smartphones Sensors in Daily Life Activities
The marvelous expansion of sensory technology in
smartphone has enabled to track dynamic information about
environmental impacts (e.g., noise level, air pollution level,
humidity, and temperature etc.), and objects movements
patterns (e.g., people's activities, and traffic and road
conditions etc.) etc., and model them in fruitful ways (e.g.,
rendering of tracking information on a map and sharing
users’ contextual information with online social
communities etc.). In addition to using smartphone sensing
capability to solve daily life problems, sensory applications
could also ease quick data gathering in an urgent situation
such as during disaster-relief operation (i.e., earthquake,
flood, or outbreak of a disease etc.) personnel (e.g.,
sociologist, engineers, doctors, biologists, aid-worker etc.)
can use their smartphones to sense, monitor, and visualize
real world phenomena for realizing public-health threats, and
environmental hazards etc. This growing interest in
smartphone sensing is due to the technological
advancements [24]. First, the availability of cheap embedded
sensors in smartphone has made possible the creation of
disruptive sensing applications. Second, smartphones are
open and programmable which eliminates the barriers of
entry for third-party programmers. Third, vendors have app
stores allowing application developers to deliver their
applications to large number of user across the globe. Fourth,
developers can use the high valued resources and services on
back-end servers of cloud computing for computation of
large scale sensory data and other advanced processing.
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 11 ISSN 1999-4974
A number of real world scenarios can be outlined
utilizing smartphone sensing capabilities. An excerpt of
possible applications of smartphone general purpose sensors
in users' daily life activities are summarized in Table II.
A context aware smartphone can recognize the context
of a user using sensors and can either change its behavior
accordingly or initiate a service automatically. For
example, a smartphone might either not accept any call
or switch off entirely in situations when a user is in
bathroom or in meeting etc.
Smartphone can automatically obtain weather
information (e.g., temperature, humidity, and wind force
etc.) either using embedded sensors or nearby connected
external sensors and throw an automatic text message to
formers using an automatic notifications application to
inform them about potential dangers to their seeds or
crops in advance.
Smartphone can use sensing capabilities for accurate
capturing of information about traffic and road
conditions and share them with other people in an area
using some wireless networking technology (e.g.,
Bluetooth, GSM network, Wi-Fi etc.) to help them in
finding alternative and time saving paths to their
destinations.
Smartphone can use embedded sensors or external
sensors attached to different body parts of a person to
get health information (e.g., measuring blood pressure,
heart beat, temperature level, and obesity etc.) in real-
time and either prompt messages to the users or inform
emergency responders to take appropriate actions. For
example, a smartphone sensing system might observe a
person’s food intake, calculate the amount of calories
taken, and suggest him the amount of exercises he is
needed to burn extra calories.
Having real-time knowledge of altitude value and
turning GPS on and off accordingly can be preemptive
to a bad situation for hikers in a mountainous region. An
application using altimeter sensor can trigger an alarm
reminding the altitude level upon reaching a threshold
elevation value and might turn on the GPS upon
reaching a threshold elevation value, saving battery
power considerably while recording tracks relatively
accurately.
C. Smartphones Sensors and Battery Power
The limited battery capacity of smartphone can hinder
and restrict the effectiveness of sensors-based applications
and services irrespective of their usefulness. Among the
others, noticeably the embedded sensors in smartphone are
the major sources of battery power consumption. For
example, Nokia 95 smartphone can support telephone
conversation for more than ten hours if battery is fully
charged, but a turned-on GPS receiver can completely drain
out the same battery within six hours whether getting GPS
readings or not [1]. However, sensors vary in battery power
consumption rates where some are very greedy as compared
to others. For example, a switched on GPS receiver can
completely drain a Nokia N95 8GB battery in 7.1 hours and
11.6 hours respectively in indoor and outdoor, whereas,
accelerometer can took 45.9 hours to completely drain out
the same battery [25]. Typically, the energy consumption
rate of a sensor depends on its sampling rate for reading
contextual data: the higher the sampling rate the higher the
energy consumption and vice versa. For example,
accelerometer, gyroscope, barometer, and magnetometer
sensors reads contextual data on uniformed sampling rate,
whereas, proximity, and ambient light sensors reads
contextual data on non-uniformed sampling rate.
In addition to sensors, the integration of diverse
functionalities such as voice communication, web browsing,
audio and video playback, SMS/EMS and email
communication, and gaming etc., can also produce sever
pressures on battery lifetime. The application developers are
providing sophisticated solutions and engaging usage
experiences through applications by exploiting the desktop-
like features of smartphone such as powerful processor,
RAM, sensors, and bright colorful display. But, being power
hungry, the continuous usage of these hardware components
can shorten the battery life significantly. An investigation
has shown that majority of Android applications have been
reported suffering with energy inefficiency problems by the
users. Most of the problems are caused by sensor for two
reasons. First, the Android framework gives full sensor
management control to developers, which could result into
excessive power wastage if mismanaged. Second, most of
the Android applications are developed by small teams
tending to provide functionalities without dedicated quality
assurance and majorly overlook power inefficiency
problems.
Under these circumstances, an effective power
management is needed intensively. An effective and efficient
power management is subjected to clear understanding of
where and how power usage formula. It should be defined
that which part of a system should use how much of the
system’s power and under what circumstances [3]. In
smartphone sensing applications, power saving can be
achieved by shutting down unnecessary sensors as well as
carefully selecting sensors duty cycles (i.e., sensors will
adopt periodic sensing and sleeping instead of being sampled
continuously) [1]. Sensors sampling rate should be adjusted
according to users’ contexts. For example, GPS receiver
should be turned on while operating outdoor and should be
turned off while operating indoor. Furthermore, time
intervals should be introduced between consecutive samples.
IV. RELATED WORK
Several researchers have attempted to find out how
power is consumed in smartphone. Many researchers have
concluded power consumption as the primary problem in
smartphone management and devised their own ways to save
power. In the recent years, researchers have contributed fair
amount of work investigating smartphone applications and
services utilizing sensors data. Applications ineffectively
using smartphone sensors are explicitly found wasting most
of the energy. Most of the researchers have used a single
sensor in a big list of available smartphone sensors for energy
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 12 ISSN 1999-4974
consumption estimations. However, some of the researchers
have concluded energy consumption estimations using all of
the available smartphone sensors but they are suffering with
certain limitations as well.
Fehmi Ben Abdesslem, et al. [25] have presented
SenseLess system which leveraged the different energy
characteristics of sensors for maximizing battery life for
smartphone sensing applications usage. Each sensors (i.e.,
GPS both indoor and outdoor, microphone, Bluetooth, and
accelerometer) is used explicitly and continuously on a
Nokia N95 8GB smartphone until the battery is completely
depleted. It is found that GPS is more power hungry and
accelerometer is less power hungry among the all. The
approximate battery life for GPS (outdoor), GPS (indoor),
microphone, Bluetooth, and accelerometer is found 7.1, 11.6,
13.6, 21.0, 45.9 hours respectively. The approximate battery
life when all of the sensors are turned off is 170.6 hours.
However, SenseLess suffers from certain limitations. First, it
did not experiment other available sensors in smartphones
such as proximity sensor, light sensor, magnetic field sensor,
and orientation sensor etc. Second, it only tested power
consumption of GPS in indoor and outdoor, whereas, other
sensors power consumptions in indoor and outdoor activities
are completely ignored.
Fangwei Ding, et al. [26] have developed Android based
smart energy monitoring system SEMO for profiling
smartphone applications with battery consumption. SEMO
system works by checking the battery’s status, collecting
energy consumption data of applications in accordance to
data collection, and ranking the applications using energy
consumption rates. However, SEMO focus on recording and
understanding applications' energy consumption information
from developers' perspectives and does not record energy
consumption information of energy hungry smartphone
components such as screen light, network interfaces (e.g.,
Wi-Fi etc.), and sensors (e.g., GPS etc.).
Mian Dong, et al. [27] have described a self-modeling
paradigm namely Sesame which leverages smart battery
interface for self-power measurement without any external
assistance and gains accuracy and rate much higher than
smart battery interface using a suite of novel techniques. The
experimental result showed that Sesame generated system
energy model has 95% accuracy. They highlighted the
dependency of energy model on hardware configuration,
usage, and smartphone. After experiments, they proposed
that increase in the memory size and CPU cycles will have
effect on battery consumption. Furthermore, media player
application has been found more energy consuming
application as compared to others.
V. MATERIALS AND METHODOLOGY
To calculate and analyze the power consumption rates
of the smartphone sensors, we have implemented an Android
app namely Energy Monitoring System for Smartphone
Sensors (EM3S). EM3S can find the energy consumption of
each available smartphone sensor explicitly and in collection
with other sensors at the same time in different real world
scenarios. The considered scenarios are four in numbers
where each scenario is composed of user states (i.e., motion
or stationary), smartphone states (i.e., motion or stationary),
sensors states (i.e., on or off), environment states (i.e.,
building/indoor or open ground/outdoor), and user activities
(i.e., walking, upstairs, down stairs, standing, or sitting).
Table III depicts the four scenarios along with their
compositions.
These compositions are inspired of the real world
situations which are experienced by the users in their daily
Table III Scenarios and Their Compositions.
User
Smartphone
Sensor
Environment
Activities
Walking
Up Stairs
Down
Stairs
Standing
Sitting
Motion
Motion
ON
Building
Stationary
Stationary
ON
Building
Motion
Motion
ON
Open Ground
Stationary
Stationary
ON
Open Ground
Table II. An Excerpt of Smartphone Sensors Applications in Daily Life Activities
No
General Purpose Sensors
Applications
1
Proximity Sensor
Detecting nearby objects in different systems such as in blind people guidance systems to help them during
walking etc.
2
Accelerometer Sensor
Measuring movements, angles, inclination, and acceleration information of users while conducting a multitude
of physical activities in different systems such as old people health care systems, automatic traffic accident
detection systems, and games etc.
3
Gyroscope Sensor
4
CMOS Camera Sensors
Taking pictures of users and surrounding environment which could be used in a number of systems such as
recognizing user surrounding environment and location systems, and recognizing users' inclination systems etc.
5
Ambient Light Sensor
Measuring light intensity data of surrounding environment to be used in environmental pollution monitoring
systems, picture capturing systems, and weather forecasting systems etc.
6
GPS sensor
Measuring users' locations and direction data for using in several systems such as tourists helping systems in a
new city, and soldiers helping systems in battle field or combat etc.
8
Microphone sensor
Measuring voice levels either produced by different object in smartphone’s external environment or by the user
for using in systems such as voice identification system, environmental pollution monitoring system, automatic
traffic accident detection system, and spying helping systems etc.
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 13 ISSN 1999-4974
life activities. In daily life, a user accompanied with
smartphone could be in movement or stationary, indoor or
outdoor, and involved in an activity or sitting idle. To
accurately determine sensors power consumption rates,
EM3S relies heavily on the Android's built-in modules (i.e.,
Battery States, and Sensor classes etc.). At first, EM3S check
the battery’s status (i.e., temperature level, and remaining
power) to initialize other components of the system. During
operation, EM3S queries continuously Android services for
pitching information. The collected bunch of data includes
the remaining battery’s power at the very time, names of the
running applications (i.e., including sensors) at the time, and
Fig. 2 Screen shots of EM3S main user interface.
Fig. 3 EM3S three layer architecture.
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 14 ISSN 1999-4974
the applications' total running time so far. EM3S data
analysis and the corresponding algorithms filter the retrieved
data for finding sensor's energy consumption information
and transform it into percentage for user display. EM3S main
user interface snapshots are shown in Fig. 2.
A three-layer architecture has been proposed for EM3S
consisting of user interface layer, processing layer, and
system layer as shown in Fig. 3. Each layer is composed of
several sub-components and exploits the capabilities of the
layers below. The flow of interactions and communications
between the different layers components is depicted in Fig.
4.
A. User Interface Layer
User interface is the space where interaction between
users and EM3S takes place. EM3S user interface is easy to
use, easy to understand while having lower learning curves,
having professional aesthetics, and requiring minimum steps
to obtain the desired results. User Interface layer is composed
of parent activity, providing features to invoke other
activities such as graphs etc. The parent activity layout is
composed of numerous controls (e.g., radio buttons,
checkboxes, buttons, and progress bars etc.) for providing
rich set of features and displaying information in percentage
such as sensors turning ON/OFF, smartphone modes
changing, energy consumption rates of all active
applications, active sensors, screen light, active network
interfaces, and battery remaining power etc. Buttons on the
parent activity enables users for controlling lower layers
components such as turning ON/OFF SensorApp etc.,
displaying graph activity depicting sensors power
consumption in bar chart graph, and invoking Android’s
built-in battery information service. Parent activity works as
an inspector, continuously pooling lower layers components
for required power consumption statistics and current battery
status. When the battery power reaches a critical condition
(i.e., less than 10% etc.), parent activity can also warn users
for appropriate actions.
B. Processing Layer
Processing layer is an interface between user interface
layer and system layer where all of the technical operations
would take place. Processing layer is composed of three sub-
components namely utility service, power monitor service,
and analyzer service. Utility service is a general purpose
service which provides methods to configure the app
environment. It receives configuration commands such as
turning sensors, ON/OFF, and changing modes etc. from the
user interface layer and invoke Android’s built-in modules in
the system layer to fulfill the required tasks. Power monitor
service acts as a query, filter, and recorder. Power monitor
service starts automatically with EM3S start and periodically
queries Android’s services in the system layer to retrieve
composite information including battery related information,
active sensors and modes, energy consumption information,
and other miscellaneous information. Power monitor service
filters and splits the bunch of information received from
system layer into individual information. The service, after
processing information, records information in database in
the system layer for future necessary actions. Analyzer
service could also be started by the interface layer
components and consists of three sub-components namely
analyzer, ranker, and exporter. Analyzer component
analyzes the information recorded by the power monitor
service in the database. Ranker component uses the
information produced by the analyzer component to rank
sensors by their energy consumption rates in the different
modes. Exporter component provide methods to pull the
database object (file) into PC for performing advance
powerful analysis using statistical tools. Both ranker and
exporter will make it easy to determine which sensor
consumes more power and in which mode.
C. System Layer
The lower system layer encompasses the Android’s
built-in services and libraries which are used by the EM3S.
The important built-in services and libraries at this layer
includes BatteryStates, SensorManager, and SQLLITE.
BatteryStates service provides methods to retrieve different
types of information including battery status, remaining
power in percentage, health, and temperature etc.
SensorManager service provides methods to turn required
sensors ON/OFF. Phone modes are changed by invoking the
built-in system settings services by passing appropriate
predefined constant values and other numerical values to
adjust accordingly. SQLLITE is used for creating database
to store retrieved power consumption information that are to
be used for analysis purposes. From SQLLITE, the database
file can be exported to PC, where applications such as
NavicateLite, and MS Excel etc., can be used for conducting
more powerful analysis.
VI. RESULTS AND DISCUSSION
The current version of EM3S is developed in Java and
aimed for Android based smartphones running with Ice
cream Sandwich 4.0.3 or higher. The application is mainly
tested on QMobile A12 smartphone. In order to demonstrate
the viability of EM3S, the system is tested closely in a real
world domain. A program of user tests is developed to define
activities in all of the four scenarios (as shown in Table III).
To accomplish the objectives of our study, we considered six
different activities: walking, ascending stairs, descending
stairs, running, sitting, and standing. However, the number
and intensity of the activities varies depending on scenarios
and tests. All of the activities are performed uniformly and
randomly for a period of two hours during a scenario test. To
carry out tests, three participants are given QMobile A12
smartphones and they are trained how to use EM3S
application. The participants are instructed to test a scenario
for two hours a day continuously for a week. The data
collected after performing each test is analyzed and results
are compiled. All of the tests are accomplished inside the
premises of the University of Peshawar, Pakistan.
A. General Purpose Sensors
EM3S calculates power consumption information about
general purpose sensors in real world scenarios that are
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 15 ISSN 1999-4974
described earlier. The algorithm used by EM3S for general
purpose sensors power consumption rates estimations can be
mathematically described in the following equations:
   


Equation (1) represents the power consumed by a sensor
Psensor using the power consumption information specified by
the sensor manufacturer for a unit of time. In equation (1),
Psensor is equivalent to the nominal power Ppower, and the time
unit Ttimeunit. Equation (2) represents the total power
consumed by a sensor in a scenario while the sensor is active.
In equation (2), Psensor(j) represents the total power consumed
by a sensor Psensor in a scenario j, that is equivalent to the
summation of power consumed by the sensor Psensor in
scenario time from 1 to n.
After performing all of the tests, the power
consumptions information (i.e., obtained through (1) and (2))
for each of the general purpose sensors are analyzed and
presented in Fig. 5.1, Fig. 5.2, Fig. 5.3 and Fig. 5.4
respectively for indoor stationary, indoor motion, outdoor
stationary and outdoor motion scenarios. The overall power
consumption information of the general purpose sensors in
the considered scenarios is shown in Table IV and Fig. 5.
Comparatively, it is found that GPS sensor is more power
hungry and accelerometer is least power hungry in all of the
scenarios. GPS consume more energy because of its frequent
communication with satellites to find geo-location of
smartphone. Using the information presented in Fig. 5, the
following facts are found:
Accelerometer sensor takes readings continuously with
a predefined interval time (i.e., interval time can be
changed but in this study the default interval time is
used). Accelerometer consumed less power in indoor
stationary and more power in outdoor motion.
Furthermore, accelerometer consumed slightly more
energy in outdoor as compared to indoor. It was
expected that accelerometer will consume the same
power amount in all of the cases but slight difference
was observed. However, the difference is very small and
negligible.
Table IV General Purpose Sensors Energy Consumption in Percentage.
Sensors
Indoor
Motion
Indoor
Stationary
Outdoor
Motion
Outdoor
Stationary
Accelerometer
6.35%
5.95%
8%
6.6%
Proximity
6.7%
5.9%
5.9%
6.9%
Orientation
13.23%
9.7%
5.9%
6.9%
Light
6.9%
6%
6%
6%
Magnetic Field
13%
11.4%
12.5%
12.9%
GPS
46%
45.22%
53%
51%
Fig. 4 EM3S flow chart.
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 16 ISSN 1999-4974
Magnetic field sensor takes readings continuously with
a predefined interval time (i.e., interval time can be
changed but in this study the default interval time is
used). Magnetic field consumed less power in indoor
stationary and more power in indoor moving. Like
accelerometer, magnetic field was expected to have the
same power consumption amount in all of the case but
slight difference was observed. However, the difference
is very minute and negligible.
Proximity sensor is event based sensor and takes
readings upon event occurrences. Proximity consumed
almost the same power in all of the cases. The difference
is not because of the proximity but can be attributed to
users' mistakes in frequency and timing of event
occurrences, and users' quickness in actions.
Light sensor, like proximity sensor, is event based
sensor. Light sensor consumed same power in indoor
stationary, outdoor stationary, and outdoor motion. This
is because of the fact that the light conditions remained
the same in these cases. In indoor motion, the power
consumption is slightly greater due to varying light
conditions inside a building etc.
Orientation sensor takes readings continuously with a
predefined interval time (i.e., interval time can be
changed but in this study the default interval time is
used). Orientation consumed less power in outdoor
stationary and more power in indoor motion.
Furthermore, like accelerometer, orientation consumed
somewhat how more power in outdoor as compared to
indoor. It was expected that orientation will consume the
same power amount in all of the cases but slight
difference was observed. However, the difference is
very small and negligible.
GPS sensor is event based sensor that continuously
sense but records readings upon event occurrences (i.e.,
changing location etc.). According to expectations, GPS
consumed more power in indoor as compared to
outdoor. It is already proved that GPS consumes more
battery power in indoor as compared to outdoor [25].
However, the difference observed is not as much as
claimed by [25].
B. Network Interface Sensors
Network interface sensors are commonly used for data
communications either between smartphones and LAN,
smartphones and cellular network, or between smartphones.
Network interface sensors have turned smartphones into data
centric devices. Like general purpose sensors, the ineffective
use of network interface sensors can also be a major source
of smartphone battery power loss. Using EM3S, the energy
consumption rates of network interface sensors is estimated
in real world scenarios that are described in Table III. The
algorithm used by EM3S for general purpose sensors power
consumption rates estimations can be mathematically
described as:
  
 
  
 

 
Equation (3) represents the power consumed by a
network interface sensor for original data transmission Pdata
that is equivalent to the amount of data sent Dsent and amount
of data received Dreceived in a unit of time Ttimeunit. Equation
(4) represents the power consumed by a network interface
sensor for control data transmission Pcontrol that is equivalent
to the amount of control data sent Csent and amount of control
data received Creceived in a unit of time Ttimeunit. Equation (5)
represents the amount of power consumed by a network
interface sensor Psensor in a scenario j is equivalent to the
summation of power consumed by original data Pdata and
control data Pcontrol communicated for the duration of the
scenario time from 1 to n.
Like general purpose sensors, network interface sensors
(i.e., Wi-Fi, GSM, and Bluetooth) are also tested using the
same methodology. However, for more insight results, two
different test cases are defined for carrying out tests in the
scenarios that are network interface sensors active without
data transmission, and network interface sensors active with
data transmission. In the first test case, the sensors are active
only (i.e., in connection with nearby access facility) while
having no data transmission. In the second test case, the
sensors are active as well as having data transmission. The
power consumption rates of the network interface sensors for
the scenarios in both test cases are shown in Table V and
Table VI respectively. Furthermore, their average power
consumption rates of the sensors in both of the test cases are
also depicted in Fig. 6.1, and Fig. 6.2 respectively.
Wi-Fi consumed more power as compared to other
network interface sensors in both of the test cases. Wi-
Fi consumed less power in indoor stationary and more
power in outdoor motion. More power consumption in
outdoor motion could be due to overcoming the
obstacles, high motion, and distance from nearby access
point. Wi-Fi power consumption rate in the second test
case is also greater in all scenarios than the first test case.
Obviously, it is because of the data transmission in
addition to connection. However, averagely Wi-Fi
power consumption rate is very high and could drain out
smartphone battery exponentially. Therefore, needs
improvements.
GSM consumed less power than Wi-Fi in both of the test
cases. However, GSM power consumption rate is less
than Bluetooth rate in the first test but more in the
second test. In both of the test cases, GSM consumed
less power in outdoor stationary and more power in
indoor motion. More power in indoor motion could be
due to overcoming the obstacles, movement, signal
strength, and distance from nearby BTS. Like Wi-Fi,
GSM power consumption in the second test is greater in
all scenarios than the first test case. Obviously, it is
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 17 ISSN 1999-4974
because of the data transmission in addition to
connection. However, averagely, like Wi-Fi, GSM
power consumption rate is very high and could drain out
smartphone battery exponentially. Therefore, it also
needs improvements.
Bluetooth showed uniform power consumption in all of
the scenarios in each of the test cases. Like Wi-Fi and
GSM, Bluetooth power consumption in the second test
case is greater in all of the scenarios than the first test
case. Obviously, it is because of the data transmission in
addition to connection. However, averagely, like Wi-Fi
and GSM, Bluetooth energy consumption rate is very
high and could drain out smartphone battery
exponentially. Therefore, it also needs improvements.
Table V. Energy Consumption With No Data Transmission In Percentage.
Scenarios
Wi-Fi
GSM
Bluetooth
Stationary Indoor
22%
24%
25%
Stationary Outdoor
29%
18%
25%
Moving Indoor
25%
26%
25%
Moving Outdoor
32%
20%
25%
Table VI. Energy consumption with data transmission in percentage.
Scenarios
Wi-Fi
GSM
Bluetooth
Stationary Indoor
24%
30%
28%
Stationary Outdoor
32%
25%
28%
Moving Indoor
27%
33%
28%
Moving Outdoor
37%
28%
28%
Fig. 5 General purpose sensors energy consumption rates.
Fig. 5.1. Energy consumption in indoor stationary.
Fig. 5.2. Energy consumption in indoor motion.
Fig. 5.3. Energy consumption in outdoor stationary.
Fig. 5.4. Energy consumption in outdoor motion.
Fig. 5.5. Overall energy consumption in different scenarios.
Fig. 5.1 Energy Consumption indoor stationary
Fig. 5.2 Energy Consumption indoor motion
Fig. 5.3 Energy Consumption in outdoor stationary
Fig. 5.4 Energy Consumption in outdoor motion
Fig. 5 General purpose sensors energy consumption rates
Accelerometer
Magnetic Field
Proximity
Ambient Light
Orientation
GPS
Accelerometer
Magnetic Field
Proximity
Ambient Light
Orientation
GPS
Accelerometer
Magnetic Field
Proximity
Ambient Light
Orientation
GPS
Fig. 5.5 Overall energy consumption in different scenarios
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 18 ISSN 1999-4974
VII. CONCLUSION
Smartphone capabilities and functionalities are
increased exponentially with the integration of sensors.
Today’s smartphone come with a number of high valued
embedded sensors that are having rich sensing capabilities.
Increasing incorporation of sensors in smartphone fosters the
proliferation of various sensors-based applications especially
lifelogging. Smartphone-based lifelogging applications can
leverage the smartphones sensing capabilities for capturing
variety of content and contextual data about users' daily life
activities. However, the limited battery power capacity of
smartphones restricts the scope and applications of
smartphone-based lifelogging applications due to their heavy
reliance on sensors utilization. Smartphone embedded
sensors are noticeably found as the major source of battery
power consumption. Therefore, sensors are needed to be
used cost-effectively in lifelogging applications otherwise
can drain out battery quickly. Generally, effective and
efficient power management requires detail understanding of
sensors power consumption rates to determine where and
how power usage strategy (i.e., which sensor should use how
much of the power and under what circumstances).
This paper investigates the effects of sensors usage on a
smartphone battery lifetime. It is found that smartphone
sensors are found highly power hungry and their continuous
usage for a short period of time can result in complete
depletion of battery. EM3S app is developed for helping in
sensors power consumptions estimations in daily life
activities. EM3S is an Android based application and
implemented on Android powered QMobile A12
smartphone for performing sensors power consumption tests
in a number of real world scenarios. The open nature of
Android helped us in conducting thorough analysis which are
not possible with other commercial smartphone operating
systems otherwise. The obtained results are analyzed
statistically and found that sensors not only consumes
significant portion of smartphone battery power but they also
showed significant variations in their power consumption
rates as well which are not expected ideally. Comparatively,
GPS in general purpose sensors and Wi-Fi in network
interface sensors are found most power hungry sensors.
Furthermore, it is observed that sensors power consumption
rates are not fixed and depending on the usage environment.
Sensors showed variations in their power consumption rates
in indoor and outdoor as well as stationary and motions
situations. It is also deduced that sensors which are expected
to have the same power consumption rates in all of the
possible situations due to their operating procedures resulted
into different power consumption rates. Collectively,
smartphone sensors are found more power consuming
components which could sabotage smartphone normal
functionalities. With this work, we have delivered an
automatic system with a systematic approach for finding
sensors power consumption rates which could be helpful for
laying down novel sensors power management methods in
research laboratories around the globe.
VIII. FUTURE WORK
For future work, we are interested in finding methods for
determining the power consumption rates of sensors in
proportion to the amount of information that they captured.
We are planning to uncover the amount and quality of
sensory information captured in variable sampling rates. We
are intended to investigate that how to reduce power
consumption rates of sensors by designing ambient
intelligent algorithm(s) that will dynamically determine
optimal sensors sampling rates. However, the information
captured by an optimal sampling rate should be rich enough
to define users' contexts accurately with minimum latency.
Furthermore, we are intending to commence our future
experiments by designing real world smartphone-based
lifelogging applications that would be using multitude of
sensors for deriving more accurate and widely acceptable
results.
REFERENCES
[1] Y. Wang, J. Lin, M. Annavaram, Q. A. Jacobson, J. Hong, B.
Krishnamachari, and N. Sadeh, "A Framework of Energy
Efficient Mobile Sensing for Automatic User State
Recognition," in Proceedings of the 7th International
Conference on Mobile Systems, Applications, and Services.,
Wroclaw., Poland, 2009, pp. 179-192.
[2] A. Schmidt and K. Van Laerhoven, How to build smart
appliances?, Personal Communications, IEEE, vol. 8, pp. 66-
71, 2001.
[3] A. Carroll and G. Heiser, "An Analysis of Power Consumption
in a Smartphone," in USENIX annual technical conference,
2010.
Fig. 6.1. Energy consumption without data transmission.
Fig. 6.2. Energy consumption with data transmission.
Fig. 6. Network interface sensors energy consumption rates
Fig. 6.1 Energy consumptions without data transmission
Fig. 6.2 Energy consumptions with data transmission
Stationary Indoor
Stationary Outdoor
Moving Indoor
Moving Outdoor
Stationary Indoor
Stationary Outdoor
Moving Indoor
Moving Outdoor
Bahria University Journal of Information & Communication Technologies Vol. 9, Issue 2, December 2016
Page 19 ISSN 1999-4974
[4] K. Y. Wong, "Cell phones as mobile computing devices," IT
professional, pp. 40-45, 2010.
[5] Y. He and Y. Li, "Physical activity recognition utilizing the
built-in kinematic sensors of a smartphone," International
Journal of Distributed Sensor Networks, vol. 2013, 2013.
[6] X. Bao and R. R. Choudhury, "VUPoints: collaborative
sensing and video recording through mobile phones," ACM
SIGCOMM Computer Communication Review, vol. 40, pp.
100-105, 2010.
[7] Győrbíró, N., Á. Fábián, and G. Hományi,, "An activity
recognition system for mobile phones," Mobile Networks and
Applications, vol. 14, pp. 82-91, 2009.
[8] J. White, et al., "Wreckwatch: Automatic traffic accident
detection and notification with smartphones," Mobile
Networks and Applications, vol. 16, pp. 285-303, 2011.
[9] P. Ljungstrand, "Context awareness and mobile phones,"
Personal and ubiquitous computing, vol. 5, pp. 58-61, 2001.
[10] Corral, L., Georgiev, A. B., Sillitti, A., & Succi, G, "A method
for characterizing energy consumption in Android
smartphones," in Green and Sustainable Software (GREENS),
2013 2nd International Workshop on, 2013, pp. 38-45.
[11] Liu, Y., Xu, C., & Cheung, S. C, "Where has my battery gone?
Finding sensor related energy black holes in smartphone
applications," in Pervasive Computing and Communications
(PerCom), 2013 IEEE International Conference on, 2013, pp.
2-10.
[12] R. Meier, Professional Android 2 Application Development:
Wrox Press Ltd., 2010.
[13] K. Kim and H. Cha, "WakeScope: Runtime WakeLock
anomaly management scheme for Android platform," in
Proceedings of the Eleventh ACM International Conference
on Embedded Software, 2013, p. 27.
[14] Denning, T., Andrew, A., Chaudhri, R., Hartung, C., Lester,
J., Borriello, G., & Duncan, G, "BALANCE: towards a usable
pervasive wellness application with accurate activity
inference," in Proceedings of the 10th workshop on Mobile
Computing Systems and Applications, 2009, p. 5.
[15] Sha, K., Zhan, G., Shi, W., Lumley, M., Wiholm, C., & Arnetz,
B, "SPA: a smart phone assisted chronic illness self-
management system with participatory sensing," in
Proceedings of the 2nd International Workshop on Systems
and Networking Support for Health Care and Assisted Living
Environments, 2008, p. 5.
[16] Consolvo, S., McDonald, D. W., Toscos, T., Chen, M. Y.,
Froehlich, J., Harrison, B. Klasnja, P. LaMarca, A.LeGrand,
L. Libby, R. Smith, I. &Landay, J. A., "Activity sensing in the
wild: a field trial of ubifit garden," in Proceedings of the
SIGCHI Conference on Human Factors in Computing
Systems, 2008, pp. 1797-1806.
[17] Dai, J., Bai, X., Yang, Z., Shen, Z., & Xuan, "Mobile phone-
based pervasive fall detection," Personal and ubiquitous
computing, vol. 14, pp. 633-643, 2010.
[18] Kwapisz, J. R., Weiss, G. M., & Moore, S. A., "Activity
recognition using cell phone accelerometers," ACM SigKDD
Explorations Newsletter, vol. 12, pp. 74-82, 2011.
[19] E. Kanjo, et al., "MobGeoSen: facilitating personal geosensor
data collection and visualization using mobile phones,"
Personal and ubiquitous computing, vol. 12, pp. 599-607,
2008.
[20] Kanjo, E., Benford, S., Paxton, M., Chamberlain, A., Fraser,
D. S., Woodgate, D., Crellin, D., & Woolard, A., "Nericell:
rich monitoring of road and traffic conditions using mobile
smartphones," presented at the Proceedings of the 6th ACM
conference on Embedded network sensor systems, Raleigh,
NC, USA, 2008.
[21] Miluzzo, E., Lane, N. D., Fodor, K., Peterson, R., Lu, H.,
Musolesi, M.,Eisenman. S.B., Zheng. X.,& Campbell, A. T.,
"Sensing meets mobile social networks: the design,
implementation and evaluation of the cenceme application," in
Proceedings of the 6th ACM conference on Embedded
network sensor systems, 2008, pp. 337-350.
[22] Kansal, A., Goraczko, M., & Zhao, F., "Building a sensor
network of mobile phones," in Proceedings of the 6th
international conference on Information processing in sensor
networks, 2007, pp. 547-548.
[23] S. Ali, S. Khusro, A. Rauf, and S. Mahfooz, "Sensors and
Mobile Phones: Evolution and State-of-the-Art," Pakistan
Journal of Science, vol. 66, pp. 385-399, 2014.
[24] Lane, N. D., Miluzzo, E., Lu, H., Peebles, D., Choudhury, T.,
& Campbell, A. T. , "A survey of mobile phone sensing,"
Communications Magazine, IEEE, vol. 48, pp. 140-150, 2010.
[25] Ben Abdesslem, F., Phillips, A., & Henderson, T., "Less is
more: energy-efficient mobile sensing with senseless," in
Proceedings of the 1st ACM workshop on Networking,
systems, and applications for mobile handhelds, 2009, pp. 61-
62.
[26] Ding, F., Xia, F., Zhang, W., Zhao, X., & Ma, C., "Monitoring
energy consumption of smartphones," in Internet of Things
(iThings/CPSCom), 2011 International Conference on and 4th
International Conference on Cyber, Physical and Social
Computing, 2011, pp. 610-613.
[27] M. Dong and L. Zhong, "Self-constructive high-rate system
energy modeling for battery-powered mobile systems," in
Proceedings of the 9th international conference on Mobile
systems, applications, and services, 2011, pp. 335-348.
  • Article
    Full-text available
    A wide range of localization techniques has been proposed recently that leverage smartphone sensors. Context awareness serves as the backbone of these localization techniques, which helps them to shift the localization technologies to improve efficiency and energy utilization. Indoor-outdoor (IO) context sensing plays a vital role for such systems, which serve both indoor and outdoor localization. IO systems work with collaborative technologies including the Global Positioning System (GPS), cellular tower signals, Wi-Fi, Bluetooth and a variety of smartphone sensors. GPS- and Wi-Fi-based systems are power hungry, and their accuracy is severed by limiting factors like multipath, shadowing, etc. On the other hand, various built-in smartphone sensors can be deployed for environmental sensing. Although these sensors can play a crucial role, yet they are very less studied. This research aims at investigating the use of ambient magnetic field data alone from a smartphone for IO detection. The research first investigates the feasibility of utilizing magnetic field data alone for IO detection and then extracts different features suitable for IO detection to be used in machine learning-based classifiers to discriminate between indoor and outdoor environments. The experiments are performed at three different places including a subway station, a shopping mall and Yeungnam University (YU), Korea. The training data are collected from one spot of the campus, and testing is performed with data from various locations of the above-mentioned places. The experiment involves Samsung Galaxy S8, LG G6 and Samsung Galaxy Round smartphones. The results show that the magnetic data from smartphone magnetic sensor embody enough information and can discriminate the indoor environment from the outdoor environment. Naive Bayes (NB) outperforms with a classification accuracy of 83.26%, as against Support vector machines (SVM), random induction (RI), gradient boosting machines (GBM), random forest (RF), k-nearest neighbor (kNN) and decision trees (DT), whose accuracies are 67.21%, 73.38%, 73.40%, 78.59%, 69.53% and 68.60%, respectively. kNN, SVM and DT do not perform well when noisy data are used for classification. Additionally, other dynamic scenarios affect the attitude of magnetic data and degrade the performance of SVM, RI and GBM. NB and RF prove to be more noise tolerant and environment adaptable and perform very well in dynamic scenarios. Keeping in view the performance of these classifiers, an ensemble-based stacking scheme is presented, which utilizes DT and RI as the base learners and naive Bayes as the ensemble classifier. This approach is able to achieve an accuracy of 85.30% using the magnetic data of the smartphone magnetic sensor. Moreover, with an increase in training data, the accuracy of the stacking scheme can be elevated by 0.83%. The performance of the proposed approach is compared with GPS-, Wi-Fi- and light sensor-based IO detection.
  • Article
    Full-text available
    Sensors can be found virtually in all fields of applications for automatic real-world data acquisition for using in plethora of decision making processes. Recent advancements have turned mobile phones from simple communication devices into complete computing devices, leveraging novel systems and services. Most of the systems have used sensors for automatic data acquisition and have gained high levels of successes. Due to the increasing use of sensors in mobile phones, today’s mobile phones come with a number of sensors while experiencing the integration of new sensors in the near future. Sensors and mobile phones integration can be attributed to plenty of reasons including easy usage, cost reductions, and high productivity etc. This paper gives a comprehensive overview of sensors integration in mobile phones, sensors classification, identifies a number of available and emerging sensors in mobile phones, and outlines a number of limitations and recommendations for the research communities to address.
  • Conference Paper
    Full-text available
    Smartphone applications have millions of users. Their energy efficiency is very important. However, we investigated 174 Android applications and found 33 of them suffering serious energy inefficiency problems. Many of these problems are due to ineffective use of sensors and their data. In this paper, we propose a novel approach to systematically diagnose energy inefficiency problems in Android applications. We derive an application execution model from Android specifications, and leverage it to realistically simulate an application's runtime behavior. Our approach can automatically analyze an application's sensory data utilization at different states, and report actionable information to help developers locate energy inefficiency problems and identify their root causes. We built a tool called GreenDroid on top of Java PathFinder and evaluated it using six popularly downloaded Android applications. GreenDroid analyzed these applications in a few minutes, and successfully located real energy inefficiency problems in them.
  • Conference Paper
    Android provides a WakeLock mechanism for application developers to ensure the proper execution of applications without having to enter the sleep state of a device. When using the WakeLock mechanism, application developers should bear the responsibility of adequately releasing the acquired lock. Otherwise, the energy will unnecessarily be wasted due to a locked application. This paper presents a scheme, called WakeScope, to handle WakeLock misuse. The scheme is designed to detect and notify of a misuse case of WakeLock handling, which may arise with an application and even with an Android runtime system, and thus provides a practical tool to prevent energy waste in mobile devices. Our experiments with real applications show that WakeScope accurately detects the misused case, with runtime overhead of approximately 1.2% in CPU usage.
  • Conference Paper
    Cellular phones and tablets are ubiquitous, with a market penetration that is counted in millions of active users and units sold. The increasing computing capabilities and strict autonomy requirements on mobile devices drive a particular concern on energy utilization and optimization of this kind of equipment. In this paper, we investigate an approach to relate the energy consumption of smartphones with the operational status of the device, surveying parameters exposed by the operating system using an Android application. Our goal is to explore the means to expand the information that may help to produce more reliable measurements that can be used in further research for designing energy optimization profiles for mobile devices and identify optimization needs.
  • Article
    Full-text available
    Physical activity (PA) recognition has recently become important in activity monitoring for the public healthcare. Although body-worn sensors are well suited to collect data on activity patterns for long periods of time, users may forget to wear special microsensors. On the contrary, more and more people take smartphone with them almost anytime. At present, most popular smartphones have three built-in kinematic sensors (triaccelerometer, gyroscope, and magnetic sensor) which could be utilized to recognize PA. This study utilized three built-in kinematic sensors in a smartphone to recognize PA and found out which features derived from the three sensors were significant to different PA. We used a combined algorithm of Fisher’s discriminant ratio criterion and 𝐽 3 criterion for feature selection. A hierarchical classifiers system including fourteen classifiers was proposed and employed to recognize fifteen activities. The optimal features derived from the built-in kinematic sensors of the smartphone were selected from 140 features. The results indicated that the accelerometer was significant to PA recognition, while gyroscope and orientation sensor were effective to recognize the change of body posture and detect falls, respectively. The total classification accuracy of 95.03% demonstrated the feasibility of utilizing the built-in kinematic sensors of the smartphone to recognize PA.
  • Article
    Full-text available
    Mobile sensing and mapping applications are becoming more prevalent because sensing hardware is becoming more portable and more affordable. However, most of the hardware uses small numbers of fixed sensors that report and share multiple sets of environmental data which raises privacy concerns. Instead, these systems can be decentralized and managed by individuals in their public and private spaces. This paper describes a robust system called MobGeoSens which enables individuals to monitor their local environment (e.g. pollution and tem-perature) and their private spaces (e.g. activities and health) by using mobile phones in their day to day life. The MobGeoSen is a combination of software components that facilitates the phone's internal sensing devices (e.g. microphone and camera) and external wireless sensors (e.g. data loggers and GPS receivers) for data collection. It also adds a new dimension of spatial localization to the data collection process and provides the user with both textual and spatial cartographic displays. While collecting the data, individuals can interactively add annotations and photos which are automatically added and integrated in the visualization file/log. This makes it easy to visualize the data, photos and annotations on a spatial and temporal visualization tool. In addition, the paper will present ways in which mobile phones can be used as noise sensors using an on-device microphone. Finally, we present our expe-riences with school children using the above mentioned system to measure their exposure to environmental pollution.
  • Article
    We present a novel system that recognizes and records the motional activities of a person using a mobile phone. Wireless sensors measuring the intensity of motions are attached to body parts of the user. Sensory data is collected by a mobile application that recognizes prelearnt activities in real-time. For efficient motion pattern recognition of gestures and postures, feed-forward backpropagation neural networks are adopted. The design and implementation of the system are presented along with the records of our experiences. Results show high recognition rates for distinguishing among six different motion patterns. The recognized activity can be used as an additional retrieval key in an extensive mobile memory recording and sharing project. Power consumption measurements of the wireless communication and the recognition algorithm are provided to characterize the resource requirements of the system.
  • Article
    Full-text available
    Mobile phones or smartphones are rapidly becoming the central computer and communication device in people's lives. Application delivery channels such as the Apple AppStore are transforming mobile phones into App Phones, capable of downloading a myriad of applications in an instant. Importantly, today's smartphones are programmable and come with a growing set of cheap powerful embedded sensors, such as an accelerometer, digital compass, gyroscope, GPS, microphone, and camera, which are enabling the emergence of personal, group, and communityscale sensing applications. We believe that sensor-equipped mobile phones will revolutionize many sectors of our economy, including business, healthcare, social networks, environmental monitoring, and transportation. In this article we survey existing mobile phone sensing algorithms, applications, and systems. We discuss the emerging sensing paradigms, and formulate an architectural framework for discussing a number of the open issues and challenges emerging in the new area of mobile phone sensing research.
  • System energy models are important for energy optimization and management in mobile systems. However, existing system energy models are built in a lab setting with the help from a second computer. Not only are they labor-intensive; but also they do not adequately account for the great diversity in the hardware and usage of mobile systems. Moreover, existing system energy models are intended for energy estimation for time intervals of one second or longer; they do not provide the required rate for fine-grain use such as per-application energy accounting. In this work, we study a self-modeling paradigm in which a mobile system automatically generates its energy model without any external assistance. Our solution, Sesame, leverages the possibility of self power measurement through the smart battery interface and employs a suite of novel techniques to achieve accuracy and rate much higher than that of the smart battery interface. We report the implementation and evaluation of Sesame on a laptop and a smartphone. The experiment results show that Sesame is able to generate system energy models of 95% accuracy at one estimation per second and of 88% accuracy at one estimation per 10ms, without any external assistance. Two five-day field studies with four laptop and four smartphone users further demonstrate the effectiveness, efficiency, and non-invasiveness of Sesame.
  • Urban sensing, participatory sensing, and user activity recog- nition can provide rich contextual information for mobile applications such as social networking and location-based services. However, continuously capturing this contextual information on mobile devices is di-cult due to battery life limitations. In this paper, we present the framework design for an Energy E-cient Mobile Sensing System (EEMSS) that powers only necessary and energy e-cient sensors and manages sensors hierarchically to recognize user state as well as detect state transitions. We also present the de- sign, implementation, and evaluation of EEMSS that auto- matically recognizes user daily activities in real time using sensors on an ofi-the-shelf high-end smart phone. Evalua- tion of EEMSS with 10 users over one week shows that it increases the smart phone's battery life by more than 75% while maintaining both high accuracy and low latency in identifying transitions between end-user activities.