Sensors: Design and Implementation of A Smart LED Lighting System Using A Self Adaptive Weighted Data Fusion Algorithm
Sensors: Design and Implementation of A Smart LED Lighting System Using A Self Adaptive Weighted Data Fusion Algorithm
Sensors: Design and Implementation of A Smart LED Lighting System Using A Self Adaptive Weighted Data Fusion Algorithm
3390/s131216915
OPEN ACCESS
sensors
ISSN 1424-8220
www.mdpi.com/journal/sensors
Article
Abstract: This work aims to develop a smart LED lighting system, which is remotely
controlled by Android apps via handheld devices, e.g., smartphones, tablets, and so forth.
The status of energy use is reflected by readings displayed on a handheld device, and it is
treated as a criterion in the lighting mode design of a system. A multimeter, a wireless light
dimmer, an IR learning remote module, etc. are connected to a server by means of RS 232/485
and a human computer interface on a touch screen. The wireless data communication is
designed to operate in compliance with the ZigBee standard, and signal processing on
sensed data is made through a self adaptive weighted data fusion algorithm. A low variation
in data fusion together with a high stability is experimentally demonstrated in this work.
The wireless light dimmer as well as the IR learning remote module can be instructed directly
by command given on the human computer interface, and the reading on a multimeter can
be displayed thereon via the server. This proposed smart LED lighting system can be
remotely controlled and self learning mode can be enabled by a single handheld device via
WiFi transmission. Hence, this proposal is validated as an approach to power monitoring
for home appliances, and is demonstrated as a digital home network in consideration of
energy efficiency.
Keywords: intelligent LED; lighting system; self-adaptive weighted; data fusion; ZigBee; IR
Sensors 2013, 13
16916
1. Introduction
As an increasingly popular issue, the field of digital home services appeals to plenty of high tech
companies. The way humans go through their daily lives in todays Hollywood films could be realized
in the very near future, one of which is the digital home network aimed at facilitating humans daily
lives. Currently, digital home network technology is being developed with focus on six aspects,
namely, central control systems, security monitoring, heath care, residence monitoring, information
appliances, and energy saving. The field of central control covers system control, management
authority, etc. security monitoring covers environment monitoring, building access control, etc. health
care covers patient location tracking, bed management in hospitals, etc. residence monitoring covers
lighting control, etc. information appliances cover home automation control, and energy saving covers
efficiency improvement, power management, etc. Currently, many companies have put a great effort
into the development of central control and information appliances, while they do not pay as much
attention to the field of health care. This study is devoted to the applications of residence monitoring
and information appliances.
There exists a wide diversity of home electronics with incompatible remote controls. The motivation
of this work is hence to develop a platform, either on a smart phone or a tablet, for interoperability
among these incompatible remote controls, such that the real time monitoring on home energy use can
be achieved, and the brightness as well as the lighting modes of a smart LED lighting system can be
switched. Smart control refers to a succession of control strategies, involving experience learning,
logic operation, adaptivity, organization, debug, and so on, and is widely applied to highly uncertain,
nonlinear, or complicated systems, which cannot be well controlled by conventional approaches.
A clear disadvantage of a conventional lighting system is that it lacks the flexibility for any
relocation of light sources, and it requires a great effort to rewire the entire system once it gets big,
e.g., in a high-rise office building, etc. These days, the instant energy use in lighting in such a high-rise
building must be monitored in real time for energy saving purposes. A smart lighting system refers to
an MCU-based system integrating automation, electronics, computer, network communication, and
many more for energy efficiency improvement. In a conventional lighting system, a light source can be
merely switched on/off manually, while, instead in a smart one, various preset lighting modes are
preloaded into the lighting system, either wired or wireless, to meet the users specific needs. Besides,
conventionally, a heavily loaded lighting system necessitates a high-capacity switch, and requires a
large volume of cables to drive a distant load. In contrast, a load is directly powered by an output
driver, meaning that there is no need to increase the power capacity of a switch when the system is
heavily loaded, and it merely requires a long signal line to drive a distant load. Furthermore, a smart
lighting system can be made dimmable and controllable by timer means. As illustrated in Figure 1,
a smart LED lighting system comprises a rectifier followed by a power factor corrector and then by a
DC/DC converter [1].
Sensors 2013, 13
16917
Figure 1. Flow chart of a smart LED lighting system.
AC
Mains
AC
Rectifier
PFC
Stage
DC/DC
Stage
LED String
Control
Stages
As a rule, there are two approaches to energy efficient lighting, namely, the use of high efficiency
light sources, and the development of smart lighting techniques. An illustration of the latter is the
thermal infrared sensing technique, by use of which indoor lights can be switched on/off automatically
when there is somebody/nobody present. On top of that, a lighting system can be made adaptive, such
that the indoor brightness can be maintained at a constant level taking into account the contribution of
outdoor sunshine. As indicated by statistics, lighting, air conditioning and the rest account for 33%,
50% and 17% of energy consumption, respectively. Since the late 1960s and early 1970s, developed
countries started to develop green lighting technologies for ecological concerns.
A great challenge to be faced is the electrical wiring problem when try to build an energy efficient
lighting system in an old building. Is there a way to get the job done, but not to rewire the whole house?
The answer is affirmative. A solution to this problem is the use of short range wireless communication
techniques, namely, Bluetooth, IEEE 802.11 WiFi and infrared. For instance, the residence lighting
can be controlled by an IR remote control. There are multiple remote controls in most residences, and a
universal remote control is a must such that any of the home appliances can be controlled by such
single piece [2].
A wide variety of sensors, including IR, ultrasonic, light, illumination, voice, and Hall sensors, can
be integrated into an MCU-based LED lighting system. In this manner, various types of detected
signals can be processed in such a way that an LED lighting system can be operated in a smart way.
Due to the very weak sensed signals, the front end electronics and signal processing is seen as required
and involves an ADC(s), an MUX, a PGA, a voltage reference, an excitation source, embedded micro
processor (MCU, ASIC), memory (RAM, E2PROM), etc. Smart signal processing can be automatically
performed on the collected data by either an MCU or an ASIC inside commercial products.
As illustrated in Figure 2, a number of wireless communication modules, including ZigBee, WLAN,
TCP/IP, WiFi, etc. can be integrated into a remote control smart LED lighting system.
Figure 2. Operation sequences of a smart LED lighting system.
Sensors
Signal processing
Infrared
Intelligent processing
Control execution
Infrared
ADC
Ultrasound
Ultrasound
MUX
Photosensitive
Electronic switch
Photosensitive
PGA
Illumination
Illumination
Excitations
Acoustic
Drive power
Execution system
Acoustic
Reference
Hall
LED Equipment
Control circuit
Hall
Sensors 2013, 13
16918
2. Literature Review
As surveyed in [3], current digital home applications mainly cover six aspects, i.e., central control
systems, security monitoring, heath care, residence monitoring, information appliances, energy saving.
A central control system can be applied to fridge monitoring, and remote control living room, the
security monitoring technique is applied to garage door control, car anti-theft devices, gas leakage
monitoring, shower temperature monitoring, building access control, emergence call system, fire
alarm, video surveillance systems, and many more. Conventionally, the performance of a security
monitoring system has a direct relationship with the number of security cameras installed, and there
are inevitably some blind points when videotaping. A long term video monitoring brings about a
multitude of audio and image data. As suggested in [4], a novel security monitoring system is operated
in such a way that a residence owner can be warned by a text alert automatically, and a motion activated
security camera(s) is operated only when there is something wrong therein for energy saving purposes.
Health care technology is applied to smart beds, electric adjustable beds, smart toilets, smart sofas,
smart first aid kits, etc. residence monitoring technology is applied to smart closets, smart lamps, electric
curtain control, automated temperature monitoring, etc. information appliance control is applied to
smart kitchens, digital TVs, image phones, etc. and energy saving technology is applied to energy
saving microwave ovens, electric ovens, energy efficient wastewater treatment, energy saving heaters,
thermal bathtubs, etc. As indicated in [5], facial feature extraction is performed before family members
return home, and the images of unidentified visitors can be displayed on monitors or smart phones for
security concerns. Besides, the region of interest within an image can be skillfully specified in an
attempt to reduce the incidences of false alarms. This technology can be further applied as an auxiliary
tool for parents to keep an eye on infants and children. As stated in [6,7], optical fiber technology has
been demonstrated as a very effective approach to the integration of smart TVs and many other
services. For instance, at a minimum transmission rate of 100 Mbps for home use, high speed network
services, including high definition TV (HDTV), video on demand (VOD), and the like, can be realized.
Any type of optical fiber wireless products is not available yet in market due to a lack of transmission
protocol stipulated. Other than the high spending, it requires a great effort to implement fiber to the
home projects. Currently, there have been a wide variety of technologies, including LED, CCD and
CMOS sensor technology, solar energy technology, available for digital home services.
There are a great number of problems waiting to be solved when integrating a wide diversity of
digital home services. As suggested in [8,9], a great challenge encountered is the platforms and
interoperability among various technologies, such as Ethernet, phone lines, power lines, IEEE 1394,
USB2.0, Bluetooth, infrared, 802.11a/b/g, and so forth. Another big challenge is to provide the
required network flexibility such that any extra sensors or devices can be effortlessly introduced into
an existing smart home service network. As pointed out in [9], another challenge is to find effective
ways to make new technologies accessible for seniors.
In this work, the addition of extra IR remote control home appliances into a digital home network
merely requires more IR output channels in the aspect of hardware, while additional control interfaces
must be developed in the aspect of software instead. As indicated in [1], an MCU-based LED lighting
system is operated in such a way that the aim of cost reductions and an energy saving plan can be
implemented. In an effort to extend LED lifetime, an LED light source is dimmed or even switched off
Sensors 2013, 13
16919
once the sensed operating temperature goes beyond a threshold, and is lit in a dark place. Furthermore,
a greater number of street lights can be lit automatically in rush hour traffic due to safety concern, and
adaptive brightness is enabled according to the number of people indoor. A smart LED lighting system
is proposed based on [1] toward the energy consumption reduction target.
If the LED heat has no way to exit the LED would be continuously enveloped in high temperatures.
After a period of time the heat will cause the LED to become unstable with a decrease in brightness.
Until now most LED lights were used with cooling aluminum housing material. There are three types
of aluminum heat sinks, die-cast aluminum, extruded aluminum and fin-style aluminum. They have
different degrees of cooling capacity, of course, the cost is different and so the LED lamp is priced
differently. Therefore, when you choose LED lights you should try to understand its thermal material,
as even with a low voltage LED chip heat will be released. LED heat radiating is not good in that it
reduces the LED lights life-time to only 8,00015,000 h or less. This is the most important factor
causing increased costs.
3. Design of an LED Lighting System
All the constraints must be taken into account, e.g., the operating conditions, limitations on
electrical and optical components, cost, LED driver current, and expected system lifetime, when an
LED lighting system is designed so as to meet the users specific needs, e.g., temperature ratings,
expected brightness.
There is no way that an LED lamp can be made 100% efficient due to the inevitable power loss in
the driver of the lamp. For this sake, the power loss must be taken into account in the design phase of
an LED lighting system. Typically, an LED driver is measured to have an efficiency ranging from
80% to 90%, while an LED driver with a high efficiency over 90% is a high price one. As illustrated in
Figure 3, the drivers efficiency is found as a function of load. It is noted that a load above 50% is
recommended for optimized efficiency, namely minimized cost. For indoor use, an efficiency of 87%
is highly recommended, while a lower one is recommended for outdoor use or for an extended lifetime.
Figure 3. A plot of the LED driver efficiency versus load.
Sensors 2013, 13
16920
The power loss leads to a drop in the lamp efficiency, and the number of LED lamps is determined
so as to meet the total luminous flux requirement, expressed as:
BS = BM /(GR )
(1)
where Bs and BM represent the actual luminous flux and the target value, respectively, G and R the
optical and the thermal efficiencies, respectively.
LED operating current plays a critical role in the lighting efficiency and the lifetime thereof. A rise
in the operating current brings about an enhanced output power and requires a smaller number of
LEDs, but the price paid is a degraded efficiency, a larger sized LED driver and a shorter life cycle,
since there is a high temperature drop across LED heat channels. The minimum, rather than typical,
luminous flux requirement, as specified in an LED application note, must be met in the determination
of the number of LED lamps. Accordingly, the number of LEDs, SLED, is given as:
S LED BS / BD
(2)
where BS denotes the actual luminous flux, and BD the minimum flux emanating from each LED.
Since each LED is an independent light source, an LED lamp provides a much longer life cycle
than a conventional light source, and can be integrated with novel light sources into an existing
lighting system.
Although LED is an energy efficient light source, a large amount of heat consumption is generated
in an LED lighting system when a great number of high power LEDs operate concurrently. The efficiency
of an LED is found to decrease with the operating current, namely, the operating temperature. In an
effort to operate an LED at a high efficiency, the operating temperature must be kept low for sure.
A significant heat dissipation improvement has been made using a heat sink to which a high power
LED is mounted. In contrast, the outward transfer heat generated by an LED lamp cannot be made as
efficient as in the preceding case, due to the thermally non conductive material used in a lamp case.
In this context, a steep temperature rise brings about a degraded performance and a shortened lifetime.
There is a great challenge when dealing with a theoretical thermal analysis on an LED lamp due to the
thermal convection with complicated boundary conditions and the thermal conduction across multiple
interfaces. As a matter of fact, there is no need to analyze the heat distribution in a non-equilibrium
state, since merely an equilibrium state is the issue of interest. It is an extremely difficult task to evaluate
the heat distribution in the interior of a lamp case across multiple interfaces, and what really matters in
practical application is whether the temperature falls below the rating. A solution to this problem is the
use of thermal resistance theory, an advantage over streamline thermal analysis in this case.
There is a huge difference between the operations of an LED, an incandescent lamp and a
fluorescent lamp. Both an incandescent lamp and a fluorescent lamp are devices powered by AC 220 V,
albeit a fluorescent lamp requires a rectifier circuit and a switch. Yet, an LED lamp is indeed a DC
operated device, meaning that AC 220 V must be rectified into DC in advance. A low efficient LED
driver will degrade the total efficiency of the lamp, according to which a way must be found to elevate
the drivers efficiency. In most cases, there are two ways to tune LED intensity, one of which is
achieved by the change of DC driving current, and the other is made by pulse width modulation
(PWM). The light intensity varies linearly with the driving current, until a threshold is reached. Light
efficiency degradation is seen together with a large amount of heat at high driving currents. The light
Sensors 2013, 13
16921
intensity is tuned according to the Talbot-Plateau law in a PWM scheme. Commercial LEDs are
available with recommended driver circuits meeting EMI and other safety requirements, leading to a
short design phase. Yet, there exists a problem that the drivers recommended are measured as a rule to
demonstrate an efficiency of approximately 80%, and drivers performance variation, affecting the life
cycle together with the operating temperature of an LED, appears among those provided by LED
manufacturers [10,11].
4. System Frame
As a server, XP-8000 takes charge of data access, and gets connected to a Touch Pad, a multimeter,
an IR learning remote module, a wireless lighting controller, and many more through an RS-232 or an
RS-485 interface such that a digit home is constructed.
As illustrated in Figure 4, an IR learning remote module is wired to the server, an XP-8000, through
an RS-232 interface, while the voltage/current readings on a multimeter is transmitted to the server by
means of an RS-485 interface. A light module is instructed via a wireless controller connected to and
by the server through the RS-232 interface, and the server is operated according to the command
issued by a Touch Pad through WiFi connection.
Figure 4. Devices controlled by the server in a smart lighting system.
In the system server, a particular block of memory, referred to as the Shared Memory hereafter,
is reserved for the current status storage of various electrical devices. For instance, instructions are
issued indirectly from a mobile device, e.g., a smart phone or a tablet, using WiFi technique by way of
the Shared Memory. In contrast, devices are instructed directly by a human computer interface on a
touch screen, subsequent to which the status information stored in the Shared Memory is updated.
Accordingly, the readings displayed on a multimeter can be presented on a smart phone, a tablet or a
human computer interface on a touch screen by means of the Shared Memory.
Sensors 2013, 13
16922
The XP-8000 is designed to take the readings on a multimeter using elogger, and issue command to
an IR or a light control module using C programs.
4.2. Touch Pad Touch Pad as Human Computer Interface
The touch pad is employed as a human computer interface on a touch screen for lighting mode
selection, light tuning, and the consumer electronics control, and signal reading from a multimeter,
as illustrated in Figure 6.
Figure 6. Back and front panels of a Touch Pad.
Sensors 2013, 13
16923
installation, low cost and high flexibility. With a measurement range of 60 A (10), extensible into
100 A (16) or even 200 A (24), it has been widely applied into the monitoring of building power
systems and plant facilities.
Figure 7. Photos of (a) a multimeter and (b) a clamp-on current transformer.
(a)
(b)
Sensors 2013, 13
16924
Sensors 2013, 13
16925
Figure 10. Circuit and photo of an electronic ballast.
As pictured in Figure 11, a lighting control module is comprised of a remote light dimmer followed
by an electric ballast and then by a set of energy saving light bulbs or LEDs. In this context, wireless
remote control light sources are provided to meet the users specific needs, including specified light
brightness or light mode control, by a system server via a remote light dimmer. A scene is lit by the
remote control light sources placed behind curtains.
Figure 11. Photos of light sources, remote light dimmers, electronic ballast and scene light
field demonstration.
The SDI (Signal Digital Interface) for the electronic dimming ballast control signal is Manchester
Encoded. As a non-polar signal, it is applicable to data transmission and synchronization. The brightness,
Sensors 2013, 13
16926
in the range of 1% to the specified level particularly for lit scenes in various modes, varies exponentially
in exactly the same way as human eyes respond to light. Another point worthy of mention is that
electric ballast can be directly switched off by a DSI for energy saving purposes [2].
High performance LED driver ICs are designed to speed up the development of smart LED-based
lighting systems. In this context, non-blinking light dimmers and even a remote control lighting system
can be implemented so as to meet the color quality requirement and raise the prices of products. The
innovative development of LED manufacturing, color and packaging techniques is seen as promising
in todays market. A major advantage of LEDs over other types of light sources is the tremendous cost
saving and a long lifetime.
LEDs are demonstrated as a revolutionary lighting device in market featuring low power
consumption as well as a long term operation. This novel type of light source can be integrated into a
long distance network or employed in high valued customized lamps. As a key component of the next
generation of light sources, a well designed LED driver, together with optic accessories and heat sink,
provides the optimized performance for an LED.
5. Development Environment
In this work, the following four pieces of application software are involved in the development of
the control interfaces and system programs.
5.1. Touch Pad Development Tool
As illustrated in Figure 12, HMIWorks is employed as the development tool to design either the
ladder diagram or the C programs in Touch Pad.
Figure 12. Demonstration of HMIWorks development tool.
Sensors 2013, 13
16927
Sensors 2013, 13
16928
Figure 14. The main window for IR-210 utility.
5.3. eLogger
eLogger is a data collection tool to program the human computer interfaces for Windows CE. NET 5.0
based PACs (WinPAC, ViewPAC) and Windows CE .NET 6.0 based PACs (XP-8000 series), such
that I/O monitoring and control systems can be constructed in a simple and highly efficient manner.
Logic control feature is provided through the Shared Memory, according to which either ISaGRAF or
VS.Net or works together with eLogger for the software development for logic controllers. Illustrated
in Figure 15 is the way that I/O modules and other register data are controlled through the Shared Memory.
Figure 15. eLogger controllable devices via Shared Memory.
Sensors 2013, 13
16929
Suppose that given an observation area, there are distinct sensed data, e.g., LED 1, LED 2, LED 3,
etc. As tabulated in Table 1, Pij , i 1,2,3...r , j 1,2,3...c , represent the data sensed about the object j by
sensor i, c the number of the objects that is an unkown, and the variances of data sensed by each node
are represented as 1 , 2 ,.. r :
r
Wi Pij
i 1
r
Wi 1,0 Wi 1
i 1
( Pj is the ultimate goal of the j-th observation)
(3)
f
W 0
i
, i 1,2,...r
r
Wi 1
i 1
(4)
Sensors 2013, 13
16930
2
i
k 1
, i 1,2...r
(5)
2
k
2
and W ( W ) 2 i 0 . The solutions W1 ,...Wr are the weights in this minimization problem:
i
i
1
f min r
1
2
i 1 i
(6)
and:
r
Pj
pij
i 1
r
(7)
i 1
2
i
2
i
P11
P21
Pr1
P12
P22
Pr2
It is revealed from Equation (7) that the ultimate goal of the observations Pj shows dependence only
on the data sensed by each sensor for a given squared error. Another point worthy of mention is that
weighted valuation and valuations in batches can be expressed in exactly the same form, with the only
difference that, in the former case, the weightings are assigned to each sensor, while, in the latter, the
mean of data is treated as the update, following which the arithmetic mean is defined as:
P
1
1 r
Pij , j 1,2...c
r i 1
(8)
2
2
and the squared error is: i r 2 i .
i 1
According to Equation (6), the squared error in a neural network based data fusion algorithm is
given as
22
1
r
i 1
2
i
12 ( 22 ) 1
r
1 r 2
1
1 r
1
) ( 2 ) 2 ( i )2 1 12 22
2 i
r i 1
r i 1
i
i 1 i
Thus, it is evident that a low variation in data fusion, i.e., a high stability, is seen in this work.
(9)
Sensors 2013, 13
16931
Sensors 2013, 13
16932
Figure 18. Lighting mode demonstration.
Figure 19. A bar diagram of the average power consumption versus the lighting mode.
Sensors 2013, 13
16933
Figure 20. Lighting mode demonstration.
Figure 21. A bar diagram of the average power consumption versus the brightness.
FR : FG : FB = 1:4.5907:0.0601
(10)
Sensors 2013, 13
16934
In this context, a mixture of 1 lm red, 4.5907 lm and 0601 lm blue light leads to 5.6508 lm white
light. As a rule, RGB are referred to as the amount of color. The flux at a certain wavelength F is
represented as:
F R( R) G(G) B( B)
(11)
The color of light can be alternatively determined by the chromaticity coordinates, the relative
values of RGB, namely. r, g, b, defined respectively as:
R
RG B
G
g
RG B
B
b
RG B
r
(12)
Tabulated in Table 2 are the RGB color specifications of a typical high power LED.
Table 2. Specifications of a high power RGB LED.
Items
Color
R
G
B
Dominant
Wavelength/nm
625
537
458
Forward
Current/mA
350
350
350
Forward
voltage/V
2.77
3.49
3.76
Flux/lm
Chroma x
Chroma y
39.6
41.6
6.5
0.700
0.223
0.149
0.299
0.714
0.031
Exhibited in Figure 22 is a RGB PWM LED driver, while instead in Figure 23 is a micro controller
based PWM driver [2].
Figure 22. RGB LED PWM driver modules [2].
DC/DC
converter
(MC34063)
Red
PWM
Red Drive
Green
PWM
Green Drive
Blue
Drive
Blue
PWM
Sensors 2013, 13
16935
Figure 23. RGB control module for an LED [2].
Power and Rectifier
AW60
Key
Detection
Converter
Red,Green,Blue
PWM
KBI
PWM
Red LED
SCI
ADC
Green
LED
Connect to PC
RS-232
Ground
I/O
Control
Universal
Input,Output
Blue
LED
In a complete closed loop color control system, color compensation must be made to yield the
intended chromatic light. The consistency in light output can be well maintained by either linear or
PWM constant current LED drivers. The choice is made considering multiple quantities of interest,
e.g., the efficiency, input voltage range, the number of LEDs, etc. There are plenty of ways to control
all the drivers output current. To begin with, a reference voltage can be generated by either a DAC or
a digital potentialmeter in such a way that the output current swing can be made up to the rated output
current. Besides, PWM signals are provided by MCU as a way to modulate the drivers output current.
A PWM controller must be operated at high switching frequency to avoid any flashing light.
The above experiment indicates a high power consistency among each of the RGB components for
identical brightness. As illustrated in Figure 24, a high consistency is seen as well for various
combinations of RGB components, namely R + G, G + B, G + B and R + G + B.
Figure 24. A bar diagram of the average power consumption in RGB components versus
the brightness.
Sensors 2013, 13
16936
Sensors 2013, 13
16937
precision, the computational complexity will be raised due to a large number of sensors. Hence, there
is a trade-off between the precision and efficiency in practical applications [17].
Figure 26. A comparison result of the relative acceptance rate versus the estimated time
between theoretical approaches and experiment.
Figure 27 illustrates the results from the computational complexity derivations for the various data
fusion. From Figure 27 it is evident that the overhead incurred by the multirate data fusion is negligible
compared to the overall savings in the time domain case. The self-adaptive weighted approach shows
the expected 21% reduction over the general data fusion. It is also very evident that this study has
excellent computational complexity efficiency between the general data fusion and standard multirate
data fusion implementation measurement [18,19].
Figure 27. Graphical comparison of computational complexity for the various data fusion methods.
Sensors 2013, 13
16938
3.
Tan, Y.K.; Huynh, T.P.; Wang, Z.Z. Smart personal sensor network control for energy saving in
DC grid powered LED lighting system. IEEE Trans. Smart Grid 2013, 4, 669676.
Park, Y.J.; Lee, M.H. Cost Effective Smart Remote Controller Based on Invisible IR-LED Using
Image Processing. In Proceedings of the 2013 IEEE International Conference on Consumer
Electronics (ICCE), Las Vegas, NV, USA, 1114 January 2013; pp. 434435.
Hwang, Z.; Uhm, Y.; Kim, Y.; Kim, G.; Park, S. Development of LED smart switch with
light-weight middleware for location-aware services in smart home. IEEE Trans. Consum.
Electron. 2010, 56, 13951402.
Sensors 2013, 13
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
16939
Zanjani, P.N.; Ghods, V.; Bahadori, M. Monitoring and Remote Sensing of the Street Lighting
System Using Computer Vision and Image Processing Techniques for the Purpose of Mechanized
Blackouts. In Proceedings of the 2012 19th International Conference on Mechatronics and
Machine Vision in Practice (M2VIP), Auckland, New Zealand, 2830 November 2012; pp. 510.
Ding, Y.S.; Jin, Y.L.; Ren, L.H.; Hao, K.R. An intelligent self-organization scheme for the
internet of things. IEEE Comput. Intell. Mag. 2013, 8, 4153.
Lee, C.K.; Li, S.N.; Hui, S.Y. A design methodology for smart LED lighting systems powered by
weakly regulated renewable power grids. IEEE Trans. Smart Grid 2011, 2, 548554.
Bhardwaj, S.; Syed, A.A.; Ozcelebi, T.; Lukkien, J. Power-managed smart lighting using a
semantic interoperability architecture. IEEE Trans. Consum. Electron. 2011, 57, 420427.
Kernen, K.; Mkinen, J.; Korhonen, P.; Juntunen, E.; Heikkinen, V.; Mkel, J. Infrared
temperature sensor system for mobile devices. Sens. Actuators A Phys. 2010, 158, 161167.
Choi, S.J.; Kim, T.H. Symmetric current-balancing circuit for LED backlight with dimming.
IEEE Trans. Ind. Electron. 2012, 59, 16981707.
Park, W.-K.; Han, I.; Park, K.-R. ZigBee based dynamic control scheme for multiple legacy IR
controllable digital consumer devices. IEEE Trans. Consum. Electron. 2007, 53, 172177.
Erol-Kantarci, M.; Kantarci, B.; Mouftah, H.T. Reliable overlay topology design for the smart
microgrid network. IEEE Netw. 2011, 25, 3843.
Thomas, C.; Balakrishnan, N. Improvement in intrusion detection with advances in sensor fusion.
IEEE Trans. Inf. Forensics Sec. 2009, 4, 542551.
Cevikalp, H.; Polikar, R. Local classifier weighting by quadratic programming. IEEE Trans.
Neur. Netw. 2008, 19, 18321838.
Zaveri, M.A.; Merchant, S.N.; Desai, U.B. Robust neural-network-based data association and
multiple model-based tracking of multiple point targets. IEEE Trans. Syst. Man Cybernet. Part C:
Appl. Rev. 2007, 37, 337351.
Park, D.-C.; Woo, Y.-J. Weighted centroid neural network for edge preserving image compression.
IEEE Trans. Neur. Netw. 2001, 12, 11341146.
Sung, W.-T.; Hsiao, C.-L. IHPG algorithm for efficient information fusion in multi-sensor
network via smoothing parameter optimization. Informatica 2013, 24, 219230.
Liu, T.H.; Yi, S.C.; Wang, X.W. A fault management protocol for low-energy and efficient
wireless sensor networks. J. Inf. Hid. Multimedia Signal Process. 2013, 4, 3445.
Chen, Y.H.; Huang, H.C. The copyright protection system for android platform. Recent Adv. Inf.
Hid. Appl. 2013, 40, 1932.
Jiang, B.; Jia, K. Semi-supervised facial expression recognition algorithm on the condition of
multi-pose. J. Inf. Hid. Multimed. Signal Process. 2013, 4, 138146.
2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).