occupancy-based energy management and … energy management and campus monitoring using wireless...

International Journal of Engineering & Technology IJET-IJENS Vol:15 No:01 56

158001-4949-IJET-IJENS © February 2015 IJENS I J E N S

Occupancy-Based Energy Management and Campus

Monitoring using Wireless Sensor Network Rashmi Priyadarshini

a, S.RN Reddy

b, R.M. Mehra

c

a. School of Engineering and Technology

Electronics and Communication Engineering

Sharda University, Greater Noida, India

[email protected] b. Department of Computer Science

IGIT, GGSIP University

Delhi, India

[email protected]. School of Engineering and Technology

Electronics and Communication Engineering

Sharda University, Greater Noida, India

[email protected]

Abstract-- The economic growth of developing countries leads

to a huge increase of energy use in buildings in large campus like

universities and schools. A major source of high electricity

consumption is careless usage of electricity. In most cases,

building appliances systems run on fixed schedules and do not

employ any control system based on detailed occupancy

information. This paper presents the design and deployment of

occupancy detection platform that can be used for accurate

occupancy detection and electrical appliances such as lights, fans,

air conditioners are controlled by considering the occupancy

information and weather conditions. With the help of passive

infrared sensor, infrared sensor, temperature sensor and

humidity sensor, the low cost and easily deployable multi-sensor

node based wireless sensor network is developed to achieve

effective utilization of electrical energy and manpower. The

system has been implemented in laboratories, class rooms and

staffrooms of Sharda University, Greater Noida, India. The

results show that approximately 10-19% power can be saved due

to implementation of this wireless sensor module. The security of

campus against intruders, moving in these areas after working

hours, is also provided by this system.

Index Term— Occupancy detection, Campus Monitoring,

Wireless Sensor Network, Zigbee Protocol, Energy Saving, Low

Cost, Multi-sensors Nodes

1. INTRODUCTION

In developing countries like India, commercial buildings are a

major source of high electricity consumption because of lack

of awareness and careless usage of electricity [1][2]. Various

researchers are focusing on making buildings more energy

efficient by including specific areas such as heating,

ventilation and air-conditioning(HVAC) [3][4], managing IT

energy consumption [5][6] and lighting [7] within buildings.

HVAC loads include mechanical equipments for combined

heating, ventilation and air conditioning which consumes

significant amount of energy. This consumption becomes

more significant if buildings are not designed using energy

efficiency factors such as dynamic windows shadings,

centralized ventilation and cooling/heating thermal

requirements.

Several models have been developed to monitor and

control electrical energy in buildings using occupancy

detection as a driving factor. Recent efforts have been to

optimize the occupancy detection techniques, so as to improve

the accuracy of estimating the occupancy. Tuan Anh Nguyen

[8] has given a meaningful survey on energy intelligent

buildings based on user activity. He has proposed future

perspective on user activity in contrast to other smart home

applications, such as medical monitoring and security system.

An energy intelligent building might not require cameras or

wearable tags that may be considered intrusive to the user.

Instead, Wireless Sensor Network (WSN) are today

considered the most promising and flexible technologies for

creating low-cost and easy-to-deploy sensor networks in

scenarios like those considered by energy intelligent buildings.

Rong-Shue Hsiao et. al [9] have proposed a robust WSN for

building lighting framework. Passive Infra-Red (PIR) and

microphone are used for occupancy detection. The occupancy

sensors used in this work are fairly coarse-grained and

inaccurate. Furthermore, these occupancy sensors are usually

limited in scope and only control lighting. This system can

control accurately if frequent conversation is possible. It is not

applicable for individual cabins of staff or faculty or officers

working calmly for a long period. Yuvraj et. al have proposed

a WSN for detecting the occupancy to control HVAC [10].

This system fails if the door of cabin was closed and main

occupant is sitting still on chair or door remains open while

room is unoccupied because occupancy detected by reed

switch depends on closing or opening of the door.

Approximately 19% power savings has been estimated by this

approach. Further, they have shown only occupancy detection

without considering the role of environmental conditions.

Elder Naghiyen et. al have reviewed occupancy

measurement techniques [11]. In this review PIR sensor,

Carbon Dioxide sensor and Device-free Localization have

been compared and have shown that PIR based occupancy

detection is cheap, low in energy consumption, easy to deploy



and operates in real-time. Padmanabh et al. have described the

use of microphone and PIR sensors to drive efficient

scheduling of conference rooms [12]. Delaney et al. have used

PIR based wireless occupancy sensors to measure wasted

energy in lighting even when there are no occupants [13].

These efforts however neither use occupancy information to

drive actual systems nor evaluate accuracy of their detection

sensor. Other methods for detecting occupancy include sonar

based methods [14] or camera based systems [15] suffers from

cost, deployment and privacy issues. Carbon dioxide- based

occupancy detection has also been proposed but the main

limitation is delay in the response of system to detect event of

incoming people [16]. Barbato et. al have focused on

algorithms supporting user profiles based on PIR sensed data

[17]. This effort is in context of home only and is not suitable

for large campus.

This work, therefore, focused on the design and

development wireless sensor network in laboratories, class

rooms and staffrooms of Sharda University for monitoring

occupancy as well as temperature, humidity and security to

provide intelligent control on appliances. The purpose of this

system was to fulfil the occupant’s requirement for comfort as

well as reduction in energy consumption during building

operations. An estimation of reduction in energy consumption

was made using occupancy detail, weather conditions, number

of occupants per floor area, construction of sensed/controlled

area such as number of windows, materials used for walls,

floors and roofs.

2. EXPERIMENTAL

The energy-use breakdown of the department of electronics

engineering (ECE) block of Sharda University is shown in

Fig.1. The breakdown of ECE block has several subsystems:

IT loads, fans, lights and HVAC. The IT loads are irrespective

of weather, climate and seasons. In the present work, IT load

were not been controlled by occupancy driven techniques.

Lighting loads consume 16% annual electricity power of the

baseline electrical usage. In India, eight out of twelve months

are hot enough to use fans. Fan loads consume 21% annual

power consumption of the baseline electrical usage. In HVAC

load all types of air conditioners, exhausts and heaters were

included. The load consumes 38% of electrical energy per

annum of the baseline electrical usage. The baseline electrical

usage for commercial buildings is defined by Uttar Pradesh

state government [18].

Fig. 1. Energy-use breakdown of Electronics Engineering Building for one year

The proposed design controls HVAC systems along with

lights and fans based on occupancy detection to get potential

power savings per annum.

The design provides an accurate estimation of occupancy and

controlling appliances of rooms of a campus to minimize the

electrical power wastage. Fig.2 shows the block diagram of

the Wireless Sensor Network for Sharda University (WSN-

SU).



Fig. 2. Block Diagram of WSN-SU

The multi sensor nodes are used to develop a reliable and low

cost WSN-SU for monitoring the occupancy, temperature,

humidity and controlling the switching ON and OFF of

electrical appliances such as lights, air conditioners, fans,

exhausts and room heaters. The system also alerts the security

by giving alarm if motion is detected in faculty rooms,

classrooms, laboratories and offices beyond working hours. It

also facilitates that occupants can change scheduled working

time if they want to do overtime or any change in predefined

working hour.

2.1. Hardware development

The WSN-SU consists of mainly two parts:

a) Sensor-actuator node

b) Base Station

a) WSN-SU sensor-actuator node

The sensor–actuator design was based on accurate and reliable

occupancy detection. False-positives (high output even no

occupancy) may lead unnecessary switching ON of the

appliances and increase energy consumption and False-

negative (low output even there is occupancy) may lead to

discomfort of occupant. The system cost was significantly

reduced as compared with commercially available wireless

motes [19] using commercially available Integrated Circuits

(ICs). The system is incrementally deployable within entire

campus, without requiring large scale modifications in the

existing wiring system. WSN-SU was designed in three

modules: First module was for occupancy detection using a

combination of sensors: PIR sensor and Infrared sensor.

Second module was for sensing of environmental conditions

using two sensors: temperature sensor and humidity sensor. To

develop an experimental test bed and to make the node of low

cost, discrete and easily available sensors/components were

used in sensor-actuator node. LM35DZ and HR202 were used

as temperature sensor and humidity sensor respectively.

LM35DZ is a precision centigrade temperature sensor which

produces the output voltage proportional to the temperature.

Table I shows the specifications of these sensors.

Table I

Specifications of the sensors.

S.N

o

Specificati

ons PIR

IR

transrecie

ver

HR20

2

LM35D

Z

1 Function

Motio

n

Detect

or

Object

Interrupt

humid

ity

sensor

Temperat

ure

Sensor

2 Operating

voltage

2.5 to

5 v

0.2 to 2.5

V 1.5 V 4 to 30 V

3 Measuring

range

0 – 9

mt 35 inch

20 to

95%R

H

-550c to

150⁰c

4 Accuracy --- ---- ±5%R

H ±1.4%

To define time schedule of sensing/control area for different

working and nonworking days, Real Time Clock (RTC) was

used in sensor node. Third Module was used to analyse the

data received from sensors and controlling appliances

according to the sensed condition. A low power and low cost

microcontroller (PIC16F877A) based on reduced instruction

set computing (RISC) architecture was used to analyse data

received from sensors and provides proper control to

appliances.

A keyboard matrix was interfaced with microcontroller to

facilitate few selected occupants to change their time schedule

as well as threshold parameters according to their need. For

reliable, low power and robust communication between node

and local base station, Zigbee protocol was used [20]. PIR

sensor, temperature sensor and humidity sensor were



connected to analog-to-digital converter (ADC) pins of

microcontroller. IR sensor does not require ADC pins, since its

output is either high or low. Real Time Clock was connected

in each node with microcontroller. Fig.3 shows the snapshot of

the circuit. The node could operate with eight PIR sensors at a

time. For faculty rooms of Sharda University, this node could

monitor six cabins.

Fig. 3. Snapshot of proposed sensor node developed for Monitoring parameters

b. Base Station

Base Station was a computer connected with Zigbee module.

Universal Serial Bus to serial port connector was used to

interface the computers and Zigbee module. At base station,

Zigbee module is configured as co-ordinator. Fig.4 shows the

received data from different nodes at base station.

Fig. 4. Received Data Display on Base Station



2.2 Working of WSN-SU

Working of system consists of three sections:

a) Occupancy Detection and Algorithm

b) Detection of Ambient conditions and Controlling appliances

c) Deployment of nodes and wireless network

a) Occupancy Detection and Algorithm

The combination of PIR and IR sensors improve the accuracy

of occupancy detection. Generally, an off-the-shelf PIR sensor

module consists of pyroelectric infrared detector with a built-in

amplifier and a digital output circuit. It has 64 detection zones

including 16 fresnel lens arrays to collect infrared radiation

from 4 quadrants of the surfaces of PIR detector. Fig. 5a-5b

shows the top view and side view of detection pattern of PIR

sensor. The detection pattern was obtained by measuring the

range of detection as a function of angular variation. PIR

sensors are most sensitive to movement perpendicular to the

direction of sensor, as movement cuts across the wedge-shaped

segment.

Fig. 5a. Top View of sensing coverage of a wall mounted PIR

Fig. 5b. Side View of sensing coverage of a wall-mounted PIR sensor

Sensitivity is lowest for movement directly towards and away

from the sensor. For individual cabin, PIR sensor was wall

mounted so that it might be most sensitive for human

movement. If occupant was seating on chair (working on

computer or studying) then PIR cannot detect, IR sensor was

mounted near chair and distance of chair was so fixed that it

could detect a person sitting on chair. Fig.6 shows the

algorithm for occupancy detection.

12m

100

⁰

Wall

Mounte

d

PIR

Sensor 3.5

m

9 mt

60

⁰

2mt

Wall

Mounted

PIR

Sensor



Fig. 6. Algorithm for occupancy detection

In classrooms and laboratories, PIR sensors were wall mounted

and IR sensors were deployed on podium. As Instructor

stopped movement and delivered lecture at podium then IR

sensor sensed the occupancy. Fig. 7 shows the deployment of

IR sensor.

Fig. 7. Sensing coverage of a desk mounter IR sensor

b) Detection of Ambient conditions and Controlling

appliances

As soon as occupancy was detected, corresponding lights were

ON. The working hours of each sensing and controlling areas

were predefined either by occupancy pattern already collected

or by time table. Before deployment of nodes, occupancy

information of classrooms, faculty cabins, Head of the

Department (HoD) cabin and laboratories were collected by

constantly monitored the hallway to record actual occupancy

for each room. The occupancy data of different rooms were

collected for one month each during teaching period and

examination period for eight working hours per day. With 25

working days in a month, the total period amount to 200 hours.

Fig.8a and 8b show the occupancy patterns for different areas

for 200 working hours which corresponds to working hours of

one month during teaching period and examination period

respectively. Occupancy data was collected per working day

since March 2013. This data was collected by continuous

monitoring the hallways.

Fig. 8a. Occupancy detail of five zones for one teaching month

Both IR and

PIR not

detected IR and

PIR

detecte

d

IR or PIR

detected

Empty:

Appliances off

Entering:

Lights on

Occupancy:

Appliances

control

initiated

Both

IR and

PIR

not

detecte

d



Fig. 8b. Occupancy detail of five zones for one examination month

The main differences in occupancy pattern during teaching

period and examination period was in the pattern of classroom,

staffrooms and laboratory. Since during examination period

most of the staff were busy in examination duties, most of the

classrooms were engaged full time for examination while

laboratories were almost free. The occupancy patterns of HoD

and researcher cabins were almost same for both the periods.

In rooms of administrative staff, working hours are to be

defined on daily basis according to their schedule. The outputs

of temperature sensor and humidity sensor were considered

only if the occupancy is detected within defined working hour

for that area. Room heaters, air conditioner, fans, exhaust fans

were controlled according to the output of temperature and

humidity sensors. Threshold of temperature depends upon the

seasons and weathers. For example, if temperature was more

than 200C, fans were ON, if it was more than 24

0C, air

conditioners were ON and if it was less than 70C, room heaters

were ON. Similarly exhaust fans were controlled by the

humidity sensors. In case of occupancy detection beyond the

limit of working hour, a buzzer blows to alert security. Fig. 9

shows the detailed algorithm of occupancy detection based

appliances control.

Fig. 9. Working Algorithm of WSN-SU

An additional facility to fix or change the working schedule as

well as the threshold parameters for controlling the appliances

was given to few selected staff as per their need using a

keyboard interfaced with the node deployed in respective

cabin. The present base station is provided with another

feature i.e. the appliances of each sensing/control area can also

be controlled through base station by feeding the desired input

from keyboard of the base station. For example

A = AC “ON”

B = AC “OFF”



C = FAN “ON”

D = FAN “OFF”

E =ALL RELAYS “ON”

F = ALL RELAYS “OFF”

These commands are required if node is not working properly

or unwanted occupancy is detected in sensing/control area

within working hour.

c) Deployment of nodes and wireless network

The network topology and the physical location of sensor

nodes strongly influence the performance of the network from

perspective of reliable communication. Each sensor node was

connected to cluster head. Cluster heads were connected with

routers and routers were connected to the base station. The

network was self- organizing. WSN was on one of the floors of

ECE block which is composed of several faculty cabins, six

classrooms, one laboratory and a cabin of HoD. Fig.10a and

10b show the sensor node deployment and the network

topology respectively.

Fig. 10a. Building Floor Plan and Deployment of sensor nodes

Fig. 10b. Hierarchical cluster-based network topology



3. RESULTS

3.1. Occupancy Detection Accuracy

In order to test the accuracy of the system, the situations of

entering and leaving the area same as occupants would behave

normally were simulated. In faculty and HoD rooms, the

experiment was divided into two parts. During the first part,

persons entered the room, walked around and came out.

Detection of WSN-SU perfectly matched the actual occupancy.

In second part, persons entered the room and after few minutes

of walking, sat on the chair with minimal movements. IR

sensor was mounted on desk at a certain distance from the

chair, up to a maximum 52 inches, so that it certainly senses a

person sitting on the chair. Again occupancy detection of the

node perfectly matched the actual occupancy. In classrooms

and laboratories, same types of experiments were performed

and accurate occupancy detection was obtained by the WSN-

SU system. Figs.11a-11c show experimental results of

laboratory occupancy test, staff cabin occupancy test and

classroom occupancy test respectively. Accuracy of occupancy

test was ranging from 95 to 98%. The accuracy test was done

using t-test of actual occupancy data and WSN-SU occupancy

data.

Fig. 11a. Accuracy of Occupancy detection in laboratory of WSN-SU vs. Actual occupancy



Fig. 11b. Accuracy of Occupancy detection in staff cabin of WSN-SU vs. Actual occupancy

Fig. 11c. Accuracy of Occupancy detection in classroom of WSN-SU vs. Actual occupancy b) Appliances control and energy calculation



To evaluate energy efficiency, electrical energy consumption

in different zones prior to the deployment of nodes was

recorded in faculty cabins, classrooms, laboratories and HoD

cabin. Sub meters were connected to measure the power

consumption of each remote area. The power consumption

would depend on seasons (summer, winter, spring and rainy

season), daily weather, construction of sensed area (sun

facings, number of windows, material type of flooring etc.),

number of occupants in the classrooms, cabins, laboratories

etc. Electrical power consumption per month was continuously

monitored since March 2013. Analysis of the collected data

revealed that in each season, two continuous months have

almost same power consumption. Those months were March-

April, May-June, July-August and December-January. Table II

shows the parameters for occupancy, temperature, humidity

and electrical loads for two such continuous months having

almost same power consumption.

Table II

Parameters for occupancy, temperature, humidity and electrical loads for two continuous months having almost same power consumption

It may be mentioned that power consumption in the month of

September was almost same as that in the month of July or

August. Similar consumption pattern was observed for October

and April, November and March and January and February.

The number of occupants in classrooms and laboratories and

workload on faculties were almost same in the month of March

and April because these were teaching months. In the months

of May and June the occupancy pattern of classrooms,

laboratories, faculty cabins were different since these were

examination months. July and August were again teaching

months and has occupancy pattern similar to March and April.

January and December were months for examination. For all

these combinations of months, for first month the power

consumption was measured without activation of sensor-

actuator node while for second month, measurements were

done with sensor-actuator node. The consumption of power

was measured with the help of sub meters connected into

sensed/controlled areas. These experiments were conducted in

five zones: Faculty cabins, HoD cabin, Researcher’s cabin,

classrooms and laboratories since these zones are different in

terms of number of occupants, occupancy patterns and

electrical loads. Figs. 12-15 show the power consumption as

function of time with and without WSN-SU in the case of

summer, winter, spring and rainy seasons.

Months Period Season Temperature

range 0C

Relative

Humidity

range

%

Appliances to be

controlled

March-

April

Teaching Spring 13 to 25 25 to 30 Lighting loads, fans

May- June Examination Summer 25 to 45 27 to 45 Lighting loads, fans,

air conditioners

July-

August

Teaching Rainy 26 to 38 75 to 80 Lighting loads, fans,

airconditioners,

exhaust fans

September Teaching Rainy 26 to 38 75 to 80 Lighting loads, fans,

airconditioners,

exhaust fans

October-

November

Teaching Post

monsoon

20 to 25 25 to 30 Lighting loads, fans

December-

January

Examination Winter 5 to 15 50 to 55 Lighting loads, room

heaters



Fig. 12. Power Consumed with and without WSN-SU in May-June

Fig. 13. Power Consumed with and without WSN-SU in December-January



Fig. 14. Power Consumed with and without WSN-SU in March-April

Fig. 15. Power Consumed with and without WSN-SU in July-August

It is evident that the rate of consumption of power is much

lower when the WSN-SU was activated. Fig. 12 reveals that in

summer the power saving on daily basis varies from 11% to

36%m with the use of WSN-SU. The average power savings is

21%. Fig. 13 shows power consumption rate decrements in

winters with the application of WSN-SU. The average power

savings was ~ 30% while on daily basis percentage power

saving ranges from11% to 40%. Figs. 14 and 15 show the

effect of WSN-SU on power consumption in spring and rainy

seasons respectively. In spring season average power saving

was ~23% and daily power savings ranges from 14 to 38%. In

rainy season average power saving was ~ 33% while daily

power savings ranges from 11 to 39%. Therefore, it was

revealed that with the implementation of WSN-SU a power

savings of ~21 to 33% was achieved in a year.

C. Simulation

The impact of WSN-SU on energy cost was studied using

energy plus building energy simulation program [21]

developed by US department of energy. The building model

was simulated for one year. Fig. 16 shows that the monthly

savings using our occupancy systems is between 10 to 19%.



Table III shows the parameter considered for simulation in energy plus. The annual consumption of Electronics Engineering

block is 472662 kWh. At least 10% of power saving can save 47266.2 kWh per annum. Table III

Information used for the simulations

Names of Zones Classroom Staff cabin HoD cabin Researcher’s

cabin

Laboratory

Floor area(m2) 98.12 42.37 98.12 42.37 183.59

Period of

simulation

January 1 to December 31

Weather data file IND_NewDelhi.421820_ISHRAE

People Schedule Occupancy data collected before deployment(Fig. 8)

Working Hours 8:00-17:00

Cooling Set point

temperature

240C for working hours

D. Effective cost Calculation

The cost of developed node was approximately US $60 which

was less than commercially available nodes [19]. Five sensor-

actuator nodes can control 30 cabins. In terms of sensing and

controlling (occupancy detection, temperature and humidity

detection and controlling of appliances) per cabin, the cost per

node is only US$10 (US$ 60/5) which is quiet cheaper than

other proposed circuits [10]. In classrooms of seating capacity

of maximum 60 students at a time, single node is enough to

detect occupancy in different rows and controlling of 12

lights, 10 fans and at least two air conditioners. Since single

sensor-actuator node can operate with multiple sensors and

control different appliances, effective cost is much lesser than

other proposed nodes.Power consumption of node was 1.8W

using sleep mode configuration with internal wake-up using

watchdog timer. The power consumption of circuit was

calculated using OrCAD (a proprietary software tool suite

used primarily for electronic design automation) and finally

verified using multimeter.

4. CONCLUSION

The study revealed that accurate occupancy information along

with environmental conditions is critical for increasing energy

efficiency. A reliable communication was achieved between

sensor nodes and base station using Zigbee protocol. False

occupancy detection was improved using combination of PIR

and IR sensors. The use of multi-sensor nodes in WSN-SU

resulted in a cost effective and energy saving campus

monitoring system. A power savings ~10 to 19% was achieved

by the system along with the saving of manpower. The system

also provides security against theft after working hours.

ACKNOWLEDGMENT

We wish to thank Mr. Manish Shekhar, Mr. Navin Singh and

Mr. Saurabh Sinha, Graduate Engineers, Sharda University, for

sensor node development and deployment. We also wish to

acknowledge the constant inspiration and support of Dean,

School of Engineering and Technology, Sharda University.

REFERENCES

[1] Inderjeet Singh, Axel Michaelowa, Indian Urban Building Sector:

CDM Potential through Engergy Efficiency in Electricity Consumption, HWWA DISCUSSION PAPER289, Hamburgisches

Welt-Wirtschafts-Archiv (HWWA), Hamburg Institute of

International Economics,2004,ISSN 1616-4814 [2] WBCSD, Transforming the Market: Energy Efficiency in

Buildings, Survey report, The World Business Council for

Sustainable Development, April 2009.



[3] V. L. Erickson, Y. Lin, A. Kamthe, R. Brahme, A. Surana, A. E.

Cerpa,M. D. Sohn, and S. Narayanan, Energy efficient building environment control strategies using real-time occupancy

measurements, Proceedings of the First ACM Workshop on

Embedded Sensing Systems for Energy-Efficiency in Buildings, pages 19–24, 2009.

[4] G. Gao and K. Whitehouse, The self-programming thermostat:

optimizing setback schedules based on home occupancy patterns, Proceedings of the First ACM Workshop on Embedded Sensing

Systems for Energy-Efficiency in Buildings, pages 67–72, 2009

[5] Y. Agarwal, S. Savage, and R. Gupta, Sleep Server: A Software-Only Approach for Reducing the Energy Consumption of PCs

within Enterprise Environments, Proceedings of USENIX Annual

Technical Symposium (USENIX ATC ’10), 2010. [6] Y. Agarwal, S. Hodges, R. Chandra, J. Scott, P. Bahl, and R.

Gupta. Somniloquy: Augmenting Network Interfaces to Reduce

PC Energy Usage, Proceedings of USENIX Symposium on Networked Systems Design and Implementation (NSDI ’09), 2009

[7] D. T. Delaney, G. M. P. O’Hare, and A. G. Ruzzelli, Evaluation of

energy-efficiency in lighting systems using sensor networks, Proceedings of the First ACM Workshop on Embedded Sensing

Systems for Energy-Efficiency in Buildings, pages 61–66, 2009.

[8] Tuan Anh Nguyen, Marco Aiello, Energy intelligent buildings based on user activity: A survey 2013, Energy and Buildings

56(2013) , pp. 244-257

[9] Rong-Shue Hsiaoa, Ding-Bing Linb, Hsin-Piao Linc, Shu-Chun Chen, and Chen-Hua Chung, A Robust Occupancy-based

Building Lighting Framework using Wireless Sensor Networks, Applied Mechanics and Materials Vols. 284-287 (2013) pp 2015-

2020,Trans Tech Publications, Switzerland.

[10] Yuvraj Agarwal, Bharathan Balaji, Rajesh K. Gupta, Jacob Lyles, Michael Wei, Thomas Weng, Occupancy-Driven Energy

Management for Smart Building Automation, BuildSys 2010

,November 2, 2010, Zurich, Switzerland [11] Eldar Naghiyev, Mark Gillott, Robin Wilson,Three unobtrusive

domestic occupancy measurement technologies under qualitative

review ,Energy and Building, Volume 69, February 2014, Pages 507-514

[12] K. Padmanabh, A. M. V, S. Sen, S. P. Katru, A. Kumar, S. P. C, S.

K.Vuppala, and S. Paul, isense: A wireless sensor network based conference room management system, Proceedings of the First

ACM Workshop on Embedded Sensing Systems for Energy-

Efficiency in Buildings, pages 37–42, 2009. [13] D. T. Delaney, G. M. P. O’Hare, and A. G. Ruzzelli, Evaluation of

energy-efficiency in lighting systems using sensor networks,

Proceedings of the First ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings, pages 61–66, 2009.

[14] S. P. Tarzia, R. P. Dick, P. A. Dinda, and G. Memik, Sonar-based

measurement of user presence and attention, UbiComp, pages 89–92, 2009.

[15] T. Teixeira and A. Savvides, Lightweight people counting and

localizing for easily deployable indoors wsns, IEEE Journal of Selected Topics in Signal Processing, 2(4):493–502, August 2008.

[16] S. Wang and X. Jin, Co 2-based occupancy detection for on-line

outdoor air flow control, Indoor and Built Environment, 7(3):165–181, 1998.

[17] A. Barbato, L. Borsani, A. Capone, and S. Melzi, Home energy

saving through a user profiling system based on wireless sensors,

Conference On Embedded Networked Sensor Systems, pages 49–

54, 2009.

[18] http://www.noidapower.com/Download/Tariff.pdf [19] Michael Johnson, et.al, A Comparative Review of Wireless Sensor

Network Mote Technologies, IEEE Sensors, art. no. 5398442 , pp.

1439-1442 [20] Varchola, Michal, and Miloš Drutarovský, ZigBee based home

automation wireless sensor network, Acta Electrotechnica et

Informatica No 7.1 (2007): 4. [21] D. Crawley, L. Lawrie, and et. al. Energyplus, a new-generation

building energy simulation program, Proceedings of the

Renewable and Advanced Energy Systems for the 21st Century, April 1999.

occupancy-based energy management and … energy management and campus monitoring using wireless...

Documents