Intelligent Luminaire based Real-time Indoor Positioning for

Assisted Living

Iuliana Marin, Maria Iuliana Bocicor and Arthur-Jozsef Molnar

SC Info World SRL, Bucharest, Romania

Keywords:

Cyber-physical System, Ambient Assisted Living, Received Signal Strength, Indoor Localisation,

Trilateration, Levenberg-Marquardt Algorithm.

Abstract:

This paper presents an experimental evaluation on the accuracy of indoor localisation. The research was

carried out as part of a European Union project targeting the creation of ICT solutions for older adult care.

Current expectation is that advances in technology will supplement the human workforce required for older

adult care, improve their quality of life and decrease healthcare expenditure. The proposed approach is imple-

mented in the form of a conﬁgurable cyber-physical system that enables indoor localization and monitoring of

older adults living at home or in residential buildings. Hardware consists of custom developed luminaires with

sensing, communication and processing capabilities. They replace the existing lighting infrastructure, do not

look out of place and are cost effective. The luminaires record the strength of a Bluetooth signal emitted by a

wearable device equipped by the monitored user. The system’s software server uses trilateration to calculate

the person’s location based on known luminaire placement and recorded signal strengths. However, multipath

fading caused by the presence of walls, furniture and other objects introduces localisation errors. Our previous

experiments showed that room-level accuracy can be achieved using software-based ﬁltering for a stationary

subject. Our current objective is to assess system accuracy in the context of a moving subject, and ascertain

whether room-level localization is feasible in real time.

1 INTRODUCTION

In recent years, the advent of Internet of Things tech-

nologies have enabled the development of smarter so-

lutions for various problems or aspects of life. Accu-

rate indoor localisation and tracking of people proves

to be one such aspect that is highly beneﬁcial for mul-

tiple purposes (Zafari et al., 2019). These include

indoor navigation within large or complex structures

such as shopping malls, airports, museums, ofﬁces or

healthcare facilities; tracking staff or mobile equip-

ment in factories or hospitals; facilitating cultural ex-

periences augmented according to visitor context in

exhibitions, museums or sports. One speciﬁc, but

highly impactful use case for accurate indoor locali-

sation is within the context of assisted living (Bianchi

et al., 2018).

The World Health Organisation expects the pop-

ulation of people over 60 to double worldwide by

2050 (World Health Organization, 2015). However,

the same report ﬁnds that the latest advances in tech-

nology and medicine have yet to be adapted to address

the problem, which is expected to impact healthcare

and local government budgets. The study by (de Mei-

jer et al., 2013) shows a 4% year over year increase in

healthcare expenditure, from which an important part

is expected to be allocated for the needs of the older

population. This creates an opportunity where recent

technological advances can be leveraged to improve

the level of care and quality of life for older adults

living within their own homes, as well as in nursing

and residential facilities.

The present paper addresses the challenge of real-

time indoor positioning in the context of the i-Light

(Marin et al., 2018) cyber-physical system for home

monitoring and assisted living. The system was de-

veloped under funding from the European Union’s

Eurostars programme and uses intelligent luminaires

that were developed as part of the project for indoor

localisation. We present an experiment in which these

luminaires are employed to obtain the real-time in-

door position of a moving person. The experiments

were performed in a three-room dwelling, with one

luminaire deployed in each of the rooms. Luminaires

incorporate a Bluetooth Low Energy (BLE) module

used both for localisation and communication.

Three scenarios were considered, in which the

person moved from one room to another, in sequence,

548

Marin, I., Bocicor, M. and Molnar, A.

Intelligent Luminaire based Real-time Indoor Positioning for Assisted Living.

DOI: 10.5220/0009578705480555

In Proceedings of the 15th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2020), pages 548-555

ISBN: 978-989-758-421-3

considering ﬁxed start and ﬁnish positions within

each room. Each scenario was repeated three times

in the exact same conditions, for veriﬁcation and re-

ﬁnement of experimental observations. Indoor local-

isation was obtained via trilateration in conjunction

with nonlinear least squares optimisation. The intelli-

gent luminaires collect the Received Signal Strength

Index (RSSI) from the monitored user’s smartphone

every 5 seconds and transfer these to the system’s

software server, where the Levenberg-Marquardt al-

gorithm (Gavin, 2011) is used to determine the loca-

tion of the monitored person.

Our initial indoor positioning experiments that de-

tail several software-based approaches to improve lo-

calisation accuracy for stationary targets are presented

in previous work (Marin et al., 2019). The main ob-

jective of our current work is to evaluate the system’s

performance with regard to indoor localisation of a

moving subject. Accurate indoor positioning of mov-

ing subjects will lower the system’s reaction times

to potentially dangerous situations, and will facilitate

the implementation of more advanced approaches for

characterising user behaviour and detect changes in-

dicative of cognitive impairment.

2 RELATED WORK

Solutions for indoor localisation and tracking have

undergone great development in recent years, par-

ticularly due to their many and diverse applications.

They are used in retail to improve customer experi-

ence, in manufacturing to increase efﬁciency by ve-

hicle tracking or staff shifting, in culture and enter-

tainment by improving visitor experience, as well as

in healthcare for staff and device tracking, or visi-

tor assistance. Although GPS technology has proven

to be quite reliable outdoors, it does not represent

a solution for indoor localisation due to poor signal

strength (Ozsoy et al., 2013). Obtaining accurate in-

door localisation is a challenging task particularly be-

cause of the multitude of obstacles in the environ-

ment, as well as reﬂection and refraction that affect

computations (Miao et al., 2018). This has led to

signiﬁcant progress in creating solutions for employ-

ing other types of technologies for indoors, such as

Wi-Fi, Bluetooth, Global System for Mobile commu-

nications (GSM), Radio Frequency Identiﬁcation, ul-

tra wide band, acoustics and optics (Ta, 2017; Lym-

beropoulos et al., 2015; Xiao and Zhou, 2016; Caron

et al., 2017).

Techniques such as Time of Flight (ToF) and Time

Difference of Arrival (TDoA) are popular when work-

ing with radio signals. ToF indicates the time a radio

signal needs to travel from a transmitter to a receiver

and back. The disadvantage of ToF is that the accu-

racy is limited by the precision of hardware timers

(Wibowo et al., 2009). ToF can be improved by us-

ing the Bluetooth standard as a channel which miti-

gates the consequences of signal fading and inference

(Giovanelli and Farella, 2018).

TDoA uses the difference given by the arrival time

of the signal sender and receiver (Yu et al., 2019). The

difference is computed as being the multiplication be-

tween the speed of light in vacuum and the time dif-

ference. If there is an obstruction along the path of the

signal, then this will result in wrong target position-

ing. Using the Angle of Arrival technique, the signal

receiver determines the angle based on the direction

of the incoming signal (Wielandt and De Strycker,

2017). Two or multiple nodes are set at known loca-

tions. The target position is given by the intersection

of the lines given by the sensing nodes. This method

is called triangulation. Infrared positioning is used

by many electronic devices, but the accuracy is low,

because rays cannot pass through walls or similar ob-

stacles (Randell and Muller, 2001). Moreover, two

infrared waves can interfere with each other. Ultra-

sound positioning is useful at night. The waves are

characterised by low penetration through obstacles.

As for infrared, a drawback is the interference of ul-

trasonic waves and reﬂection. A solution would be to

combine it with ToF (Qi and Liu, 2017).

Indoor positioning based on 3D cameras is use-

ful and precise, but it does not offer privacy for the

persons in the monitored area (Li et al., 2010). Inten-

tional blurring using the bokeh effect was evaluated

for determining indoor positioning based on the im-

ages taken with the user’s mobile phone camera (Lee

et al., 2019). Blurring was proved to enhance position

accuracy when determining the distance between a

lamp with a single LED and the user’s mobile phone.

In this way, optical-power saturation is prevented and

accuracy is enhanced. The coordinates of the person

were obtained using the received signal strength and

the angle of arrival. For the received signal strength,

the optical power is determined using the area of the

LED which appears in the photos. In the case of an-

gle of arrival, the angles are determined using the re-

ceived images, where the centre of the light coming

from the LED was analysed. Using the two methods

the 3D rectangular coordinates of the monitored per-

son were obtained.

Both industrial players and the academic commu-

nity have invested resources in solving the problem of

accurate indoor positioning. One company that lever-

ages the wide interest in indoor localisation and fo-

cuses on integrated location-based services is Navig-

Intelligent Luminaire based Real-time Indoor Positioning for Assisted Living

549

ine (Navigine, 2020). They provide an indoor posi-

tioning platform allowing other developers and sys-

tem integrators to build and create indoor navigation

and tracking services. Their platform targets mobile

developers, navigation software vendors, BLE bea-

con manufacturers, real time location systems solu-

tion providers, mapping companies, indoor naviga-

tion companies and others.

Sewio (Sewio, 2020) provides a complete plat-

form for indoor tracking based on ultra-wideband

technology. The system consists of a hardware com-

ponent that includes signal transmitters called tags,

which are attached to monitored targets, and signal

receivers called anchors. Anchors are installed within

the location and must cover the entire area for efﬁ-

cient tracking. The software solution of the platform

includes various components for deployment and con-

ﬁguration, planning, optimal anchor distribution and

type, real-time overview of the system’s performance,

remote settings updating, visualisation and analytic

tools.

The indoor tracking solution offered by WI4B

(WI4B, 2020) consists of a mesh network of nodes

deployed on location that are responsible with collect-

ing and relaying data, managing performance and ex-

changing data with a host application. Wearable tags

must be attached to the monitored persons or assets

in order to allow them to be tracked. Another com-

pany that specialises in real time indoor positioning is

Pointr (Pointr, 2020), whose solutions range from in-

door positioning and navigation to mapping, tracking

and location based analytics. The positioning system

employs beacons, inertial sensors and machine learn-

ing algorithms for both localisation and navigation.

The real time localisation systems from infsoft (inf-

soft, 2019) provide indoor navigation, geo-based as-

sistance, analytics and tracking. These solutions use

mainly Bluetooth beacons, Wi-Fi or ultra-wideband

for indoor positioning.

Extensive academic research has also been con-

ducted with regard to indoor localisation and vari-

ous solutions have been proposed. Many employ

RSSI (Luo et al., 2011; Sadowski and Spachos, 2018;

Maduskar and Tapaswi, 2017), which can be used to

determine the distance between a target and the de-

vice measuring the received signal. However, one

drawback of this method is that RSSI varies over time

due to multipath fading (Pu et al., 2011), degrading

localisation accuracy. Several approaches to allevi-

ate this problem have been proposed, such as the Si-

multaneous Localisation and Conﬁguration algorithm

proposed by (Bulten et al., 2016) as well as Kalman

ﬁlter based approaches (Robesaat et al., 2017; Sung,

2016).

(Trogh et al., 2015) propose a system based on the

Viterbi principle in conjunction with semantic data.

They start from an RSSI ﬁngerprinting technique,

consisting of an ofﬂine and an online training phase

and then employ the Viterbi algorithm to determine

the most likely sequence of positions, using the sum

of mean square errors between measurements and ref-

erence ﬁngerprints.

In contrast with the aforementioned studies and

technologies, we present a system designed for home

monitoring and assisted living, which includes an in-

door localisation component based on intelligent lu-

minaires. As opposed to other solutions for indoor

positioning, our system does not require ﬁngerprint-

ing, tags or pre-installed anchors. The central compo-

nents are a number of BLE-enabled devices that in-

clude sensing, communication and lighting modules

(Marin et al., 2018), and are deployed in the form of

intelligent luminaires.

The luminaires were created to replace existing

light bulbs and to use existing energy infrastructure,

thus permitting easy deployment, with no additional

wires or devices, while offering reduced costs. The

luminaires communicate via Wi-Fi with the system’s

server and collect localisation information from the

monitored person’s smartphone via the BLE wireless

protocol. This information is relayed to the server in

real-time, which uses trilateration and various opti-

misation algorithms to accurately compute the user’s

location.

3 REAL-TIME INDOOR

LOCALISATION WITH

INTELLIGENT LUMINAIRES

This section provides a concise description of the i-

Light cyber-physical system. The innovative aspects

of the platform hardware and software are further

detailed in our previous work (Bocicor et al., 2017;

Marin et al., 2018; Marin et al., 2019).

The platform hardware is based on luminaires that

were custom developed as part of the project. They

seamlessly replace regular light bulbs and use ex-

isting energy infrastructure. Each luminaire is ei-

ther smart, or dummy, according to its processing

and communication capabilities. Both types include

a lighting module, so they can be used as regular

light bulbs. Smart luminaires incorporate a sensing

module for monitoring ambient conditions as well as

a Wi-Fi and Bluetooth Low Energy communication

module. Dummy luminaires are smaller and more

cost-effective. Their main function is to scan the

ENASE 2020 - 15th International Conference on Evaluation of Novel Approaches to Software Engineering

550

environment for Bluetooth signals and participate in

the indoor localisation process by collecting RSSI

measurements from the monitored person’s device.

Each dummy luminaire is connected to a smart lu-

minaire, to which it sends gathered RSSI readings

via BLE. Each smart luminaire may receive measure-

ments from several dummy ones. Smart luminaires

have the capability to communicate with the system’s

server via Wi-Fi and as such are used to process and

relay aggregated measurements that include ambient

conditions and localisation data to the software server

for additional processing. The indoor positioning pro-

cess takes place on the server, where trilateration and

more advanced techniques, such as Kalman ﬁlters

(Marin et al., 2019) or optimisation algorithms are

employed. In our previous work we have concluded

that the accuracy required for the project’s purposes

can typically be achieved using up to three luminaires

per room, depending on room size and shape. Each

room should have one smart luminaire, to ensure the

stability of the connection between smart and dummy

luminaires, as well as that to the software server.

The platform includes a software server that pro-

vides a suite of custom developed, multi-platform ap-

plications. Together with the intelligent luminaire

network, they enable the system to provide conﬁg-

urable home monitoring for older adults. The sys-

tem uses ambient condition and RSSI data to provide

indoor localisation, reports and advanced visualisa-

tions. Furthermore, the system provides notiﬁcations

and real-time alerts to family members or caregivers

in case a potential danger is detected. For example,

this can happen when the monitored person is im-

mobile, but not in their bed, or when ambient con-

ditions such as temperature or volatile gas concentra-

tions reach unhealthy levels.

4 EXPERIMENTS AND RESULTS

The proposed experiment is the direct continuation of

our previous work (Marin et al., 2019), where sev-

eral software-based approaches were evaluated to im-

prove the accuracy of indoor localisation for a station-

ary subject. We extend our evaluation to a moving

subject, and study the impact of their movement on

accuracy. Furthermore, we evaluate the consistency

of recorded values with respect to errors by repeating

each scenario three times.

Experiments were carried out within an apartment

using n-dimensional space trilateration, with nonlin-

ear least squares optimisation. The algorithm’s in-

put is represented by the known positions of the lu-

minaires and the distances from the monitored target

Figure 1: Optimized Levenberg-Marquadt algorithm ﬂow

(Le Duc-Hung et al., 2012).

to each of them. These distances are computed based

on the recorded RSSI values, the path-loss exponent

and the signal strength measured at a distance of one

meter using the following lognormal model (Chern-

tanomwong and Suroso, 2011; Huang et al., 2015;

Marin et al., 2020):

d = 10

A−RSSI

10·n

(1)

The signiﬁcance of the variables in Formula 1 is as

follows:

• n represents the path-loss exponent, ranging be-

tween 2 to 6 for indoor environments.

• A is the signal strength expressed in dBm, mea-

sured at one meter. This parameter is experimen-

tally computed once for each type of luminaire.

• RSSI is the received signal strength index.

• d is the computed distance. It is the distance be-

tween the luminaire that makes the measurement

and the wearable Bluetooth device.

The output is the centroid that determines the posi-

tion of the monitored person. At least three lumi-

naires are required, along with their locations to allow

the computation of the target person’s position. The

Levenberg-Marquardt algorithm is used to solve the

nonlinear least squares problem (Gavin, 2011). It is

a combination of the gradient descent and the Gauss-

Newton methods for optimisation. Starting from an

initial guess of the parameters, it optimises them in

an iterative manner: when they are far from their op-

timal value, the algorithm behaves like gradient de-

scent, while when they are closer, it switches to em-

ploying the Gauss-Newton method.

The parameters to be optimised are represented by

the two dimensional coordinates of the monitored per-

Intelligent Luminaire based Real-time Indoor Positioning for Assisted Living

551

Figure 2: Floor plan of the dwelling used for experimenta-

tion. Evaluated scenarios were (a) walk from the bed to the

closet; (b) walk from the closet to the bed; (c) walk from

the bed to the desk chair.

son’s location. They can be determined from the fol-

lowing system of at least three

equations: (x −x

)

(y−y

)

= d

, where (x

, y

) are the coordinates of the

luminaire, i ∈ {1, 2, · ·· , n}, with n ≥ 3 and (x, y) is

the (unknown) position of the monitored person. The

objective function is thus a non-linear function of the

parameter vector P = [x, y]

and the idea is to ﬁnd

the parameters such that a minimal error between the

model and observed measurements is obtained.

The general form of the optimisation algorithm is

described in Figure 1. It starts with an initial guess for

the ﬁtting parameter values (P). In each iteration the

parameter vector is replaced by a new estimate (P +

∆), according to the Levenberg-Marquardt formulae

(Le Duc-Hung et al., 2012). This process continues

until the calculated step falls below a predeﬁned limit.

4.1 Experimental Setup

The experiment was carried out in three rooms of an

apartment located in a residential building. The layout

of the relevant rooms is illustrated in Figure 2. For the

sake of brevity, the kitchen and bathroom are not rep-

resented. Exterior and interior walls are constructed

of brick and cement. Exterior walls have a thickness

of 35cm, while the thickness of interior walls is 17cm.

As shown in Table 1, building materials have an

important deleterious effect on signal strength, mak-

ing accurate localisation based on the raw values of

signal strength difﬁcult. As shown in Figure 2, one

smart luminaire was ceiling-mounted in each room,

One equation per each luminaire.

Table 1: Absorption of radio frequencies in construction

materials relevant to our experiment (Buhagiar, 2018).

Materials Absorption Rate (dB)

Brick wall 6-15

Cement wall 4-6

Glass wall 6

Metal door 6-10

Plasterboard 3-5

Window 3

replacing regular light bulbs. For the duration of the

experiment, the luminaires had a stable Wi-Fi connec-

tion to the software server.

The experiment consisted of three scenarios that

were carried out by a healthy person using a smart-

phone as a BLE signal source. The three scenarios,

also illustrated in Figure 2 are: (a) start from the bed,

walk to the closet and stop; (b) start from the closet,

walk back to the bed and stop; (c) start from the bed,

walk to the desk chair and stop. These were carried

out in sequence, in this order, three times. Care was

taken to ensure that every time the scenario started

from the same initial conditions. Measurements were

recorded every ﬁve seconds and the periods for each

scenario were slightly different, depending on the dis-

tance the movement covered, as well as on how long

the person was static at the end of the movement

trail. The indoor locations are computed according

to Formula 1 and using the Levenberg-Marquardt al-

gorithm, by taking into consideration the last three

recorded RSSI values by each luminaire, to attenuate

single reading RSSI errors. Each scenario ends with

the person staying in the same place for an interval

between 50 to 70 seconds. This allows studying the

effect movement has on the accuracy of localisation.

It also allows evaluating how accuracy modiﬁes when

the subject ceases movement.

4.2 Results

Figure 3 illustrates the errors obtained in each of the

three tested scenarios, considering three runs for each

scenario.

In all considered cases, computed indoor positions

are coherent over repeated runs, namely the errors are

highly consistent over time, which indicates the ro-

bustness of our method. This also leads us to be-

lieve that Bluetooth ﬁngerprinting could be employed

to further improve the accuracy of localisation during

movement. We intend to further explore this in future

research.

Another important aspect targeted the impact of

movement on localisation accuracy. Results obtained

ENASE 2020 - 15th International Conference on Evaluation of Novel Approaches to Software Engineering

552

(a) Person walks from the bed to the closet and stands in

front of it.

(b) Person walks from the closet back to the bed and stands

next to it.

down.

Figure 3: The three scenarios of the experiment. Colours

synchronised with Figure 2. The horizontal axis represents

time, expressed in seconds and the vertical one is error,

measured in meters. Each scenario was played out three

times, as illustrated in the subﬁgures.

Table 2: Average errors and standard deviations (over the

three runs) obtained for all three scenarios.

Average error (m) Standard deviation (m)

Walking Standing Walking Standing

1 0.76 1.05 0.27 0.17

2 0.86 1.25 0.36 0.12

3 1.17 1.61 0.60 0.28

in all three cases indicate that the average error corre-

sponding to the period the person was stationary are

higher than when the person is moving. However,

as depicted in Table 2 as well as Figure 3 plots, the

standard deviation is lower. This indicates that the er-

rors tend to stabilise and it can be noticed that they

even slowly decrease during the time in which the tar-

get is immobile. As such, when measurements are

recorded over longer periods of time they will tend to

be more reliable and less prone to individual errors

in measurement. The experiments show that without

additional software processing (Marin et al., 2019),

the system requires at least a minute to stabilise RSSI

values, leading to more precise localisation results.

The differences in average errors between the

moving and static experiments are particularly due to

the higher errors obtained during movement, when

the person was near doors, but which are also re-

ﬂected in the error computation for the standing pe-

riod. They are inﬂuenced by readings collected during

movement, each position being computed by consid-

ering the past three RSSI values obtained from each

of the three luminaires. These higher errors can also

be observed in Figure 3 in each scenario, before the

person stops. They are most likely due to signal ab-

sorption which occurs because walls around doors use

iron-reinforced concrete pillars. While this interfer-

ence happens during movement, it has a lasting im-

pact and affects the calculations carried out when the

person is already stationary.

5 CONCLUSIONS

The presented work continues our initial evaluation

regarding the accuracy of indoor localization of sta-

tionary subjects (Marin et al., 2019). Our ﬁrst con-

clusion is that results obtained during the stationary

phase remain similar to those obtained in previous

work and provide room-level accuracy.

The second conclusion is that by itself, movement

does not decrease localization accuracy. However, ex-

cessive multipath fading can be observed when the

subject is close to architectural elements built from

materials having high absorption rates. These cause

a ﬂuctuation of signal levels that persists across sev-

eral readings, starting to stabilise once the subject is

stationary. While our previous work (Marin et al.,

2019) shows that additional ﬁltering can improve lo-

calization accuracy, these approaches must be further

adapted and evaluated in the context of a moving sub-

ject.

Future work will be carried out towards improving

the accuracy of localization via ﬁltering techniques

adapted to subject motion together with evaluating the

cost-beneﬁt of location ﬁngerprinting. Deep learning

is an additional approach that can also be used to im-

prove localisation results in an indoor noisy environ-

ment. Some experiments in this regard have already

been performed (Marin et al., 2019) for static targets,

but we plan to extend the evaluation and to also con-

sider data recorded during movement.

Intelligent Luminaire based Real-time Indoor Positioning for Assisted Living

553

ACKNOWLEDGEMENTS

This work was supported by a grant of the Ro-

manian National Authority for Scientiﬁc Research

and Innovation, CCCDI UEFISCDI, project number

46E/2015, i-Light - A pervasive home monitoring sys-

tem based on intelligent luminaires.

REFERENCES

Bianchi, V., Ciampolini, P., and De Munari, I. (2018). Rssi-

based indoor localization and identiﬁcation for zig-

bee wireless sensor networks in smart homes. IEEE

Transactions on Instrumentation and Measurement,

68(2):566–575.

Bocicor, M. I. et al. (2017). Cyber-physical system for as-

sisted living and home monitoring. In 13th IEEE In-

ternational Conference on Intelligent Computer Com-

munication and Processing, pages 487–493.

Buhagiar, J. (2018). CompTIA Network+ Review Guide:

Exam N10-007. Sybex.

Bulten, W., Rossum, A. C. V., and Haselager, W. F. G.

(2016). Human slam, indoor localisation of devices

and users. In IEEE First International Conference on

Internet-of-Things Design and Implementation.

Caron, C., Chamberland-Tremblay, D., Lapierre, C., Ha-

daya, P., Roche, S., and Saada, M. (2017). Indoor

Positioning, pages 1011–1019. Springer International

Publishing, Cham.

Cherntanomwong, P. and Suroso, D. J. (2011). Indoor local-

ization system using wireless sensor networks for sta-

tionary and moving target. In 2011 8th International

Conference on Information, Communications & Sig-

nal Processing, pages 1–5. IEEE.

de Meijer, C., Wouterse, B., Polder, J., and Koopmanschap,

M. (2013). The effect of population aging on health

expenditure growth: a critical review. European jour-

nal of ageing, 10(4):353–361.

Gavin, H. (2011). The levenberg-marquardt method for

nonlinear least squares curve-ﬁtting problems. De-

partment of Civil and Environmental Engineering,

Duke University, pages 1–15.

Giovanelli, D. and Farella, E. (2018). Rssi or time-of-ﬂight

for bluetooth low energy based localization? an exper-

imental evaluation. In 11th IFIP Wireless and Mobile

Networking Conference, pages 32–39. IEEE.

Huang, Y., Zheng, J., Xiao, Y., and Peng, M. (2015). Ro-

bust localization algorithm based on the RSSI rang-

ing scope. International Journal of Distributed Sensor

Networks, 11(2):587318.

infsoft (2019). Indoor navigation, indoor positioning, in-

door analytics and indoor tracking. https://www.

infsoft.com/. Accessed: February 2020.

Le Duc-Hung, Pham Cong-Kha, Nguyen Thi Thien Trang,

and Bui Trong Tu (2012). Parameter extraction and

optimization using levenberg-marquardt algorithm. In

2012 Fourth International Conference on Communi-

cations and Electronics (ICCE), pages 434–437.

Lee, J.-W., Kim, S.-J., and Han, S.-K. (2019). 3d visible

light indoor positioning by bokeh based optical inten-

sity measurement in smartphone camera. IEEE Ac-

cess, 7:91399–91406.

Li, X., Wang, J., Olesk, A., Knight, N., and Ding, W.

(2010). Indoor positioning within a single camera and

3d maps. pages 1 – 9.

Luo, X., O’Brien, W. J., and Julien, C. (2011). Comparative

evaluation of received signal strength index (RSSI)-

based indoor localization techniques for construction

jobsites. Adv. Eng. Inform., 25:355–363.

Lymberopoulos, D. et al. (2015). A realistic evaluation and

comparison of indoor location technologies: Experi-

ences and lessons learned. In The 14th ACM/IEEE

Conference on Information Processing in Sensor Net-

works (IPSN ’15).

Maduskar, D. and Tapaswi, S. (2017). RSSI based adap-

tive indoor location tracker. Scientiﬁc Phone Apps and

Mobile Devices, 3.

Marin, I., Bocicor, M. I., and Molnar, A. J. (2020). In-

door localization techniques within a home monitor-

ing platform. In Damiani E., Spanoudakis G., M. L.,

editor, Evaluation of Novel Approaches to Software

Engineering. ENASE 2019., volume 1172, pages 378–

401. Springer.

Marin, I., Bocicor, M. I., Molnar, A.-J., Marin, I., Vasi-

lateanu, A., Molnar, A.-J., Bocicor, M. I., Cuesta-

Frau, D., Molina-Pic

o, A., Goga, N., et al. (2019).

Indoor localisation with intelligent luminaires for

home monitoring. In Proceedings of the 14th In-

ternational Conference on Evaluation of Novel Ap-

proaches to Software Engineering, pages 464–471.

SCITEPRESS-Science and Technology Publications,

Lda.

Marin, I. et al. (2018). i-Light - Intelligent Luminaire Based

Platform for Home Monitoring and Assisted Living.

Electronics, 7(220).

Miao, Y., Wu, H., and Zhang, L. (2018). The accurate lo-

cation estimation of sensor node using received signal

strength measurements in large-scale farmland. Jour-

nal of Sensors, 2018:1–10.

Navigine (2020). Geolocation services for business. https:

//navigine.com/platform/. Accessed: February 2020.

Ozsoy, K., Bozkurt, A., and Tekin, I. (2013). Indoor po-

sitioning based on global positioning system signals.

Microwave and Optical Technology Letters, 55:1091–

1097.

Pointr (2020). Indoor positioning system. https://www.

pointr.tech/solutions/indoor-positioning-system. Ac-

cessed: February 2020.

Pu, C. C., Pu, C.-H., and Lee, H.-J. (2011). Indoor location

tracking using received signal strength indicator. In

Emerging Communications for Wireless Sensor Net-

works.

Qi, J. and Liu, G.-P. (2017). A robust high-accuracy ultra-

sound indoor positioning system based on a wireless

sensor network. Sensors, 17:2554.

Randell, C. and Muller, H. (2001). Low cost indoor posi-

tioning system. In International Conference on Ubiq-

uitous Computing, pages 42–48. Springer.

ENASE 2020 - 15th International Conference on Evaluation of Novel Approaches to Software Engineering

554

Robesaat, J., Zhang, P., Abdelaal, M., and Theel, O. (2017).

An improved ble indoor localization with kalman-

based fusion: An experimental study. Sensors, 17.

Sadowski, S. and Spachos, P. (2018). RSSI-based indoor

localization with the internet of things. In IEEE Ac-

cess.

Sewio (2020). Sewio uwb real-time loca-

tion system (rtls). https://www.sewio.net/

real-time-location-system-rtls-on-uwb/. Accessed:

February 2020.

Sung, Y. (2016). RSSI-based distance estimation frame-

work using a kalman ﬁlter for sustainable indoorcom-

puting environments. Sustainability, 8.

Ta, V.-C. (2017). Smartphone-based indoor positioning

using Wi-Fi, inertial sensors and Bluetooth. https://

tel.archives-ouvertes.fr/tel-01883828/document. Ma-

chine Learning, Universit

e Grenoble Alpes.

Trogh, J., Plets, D., Martens, L., and Joseph, W. (2015). Ad-

vanced real-time indoor tracking based on the viterbi

algorithm and semantic data. International Journal of

Distributed Sensor Networks, 11(10):271818.

WI4B (2020). Wi4b - real-time locating systems. http:

//www.wi4b.it/localization.php. Accessed: February

2020.

Wibowo, S. B., Klepal, M., and Pesch, D. (2009). Time of

ﬂight ranging using off-the-self ieee802. 11 wiﬁ tags.

In Proceedings of the International Conference on Po-

sitioning and Context-Awareness (PoCA’09).

Wielandt, S. and De Strycker, L. (2017). Indoor multi-

path assisted angle of arrival localization. Sensors,

17:2522.

World Health Organization (2015). World report on ageing

and health. https://apps.who.int/iris/bitstream/handle/

10665/186463/9789240694811 eng.pdf?sequence=1.

Accessed: February 2020.

Xiao, J. and Zhou, Z. (2016). A survey on wireless indoor

localization from the device perspective. ACM Com-

puting Surveys, 49.

Yu, D., Su, L., Zuo, Anfu, G., Sun, Z., and Wen (2019).

Time difference of arrival (tdoa) localization combin-

ing weighted least squares and ﬁreﬂy algorithm. Sen-

sors, 19:2554.

Zafari, F., Gkelias, A., and Leung, K. K. (2019). A survey

of indoor localization systems and technologies. IEEE

Communications Surveys & Tutorials, 21(3):2568–

2599.

Intelligent Luminaire based Real-time Indoor Positioning for Assisted Living

555