Design of an Energy-efﬁcient Hybrid UWB-RF Indoor Distance

Estimation System

Sabrine Khriji

, Dhouha El Houssaini

and Olfa Kanoun

Chair of Measurement and Sensor Technology, Chemnitz University of Technhology, Reichenhainer Straße 70,

09126 Chemnitz, Germany

Keywords:

WSN, UWB, RF, Indoor Localization, Energy-saving, Wireless Communication.

Abstract:

Wireless Sensor Network (WSN) plays a signiﬁcant role in modern applications by offering different services,

including state machines monitoring, predictive maintenance, and vehicle/store tracking. In such applications,

it is crucial to apply localization methods to identify the current location of sensor nodes with high accuracy.

Due to the energy constraint of sensor nodes, the energy consumption needs to be highly considered while

performing the localization process. In this regard, this paper aims to design an energy-efﬁcient and accurate

localization system. Therefore, a novel hybrid solution is designed involving the Ultra Wide Band (UWB)

technology and Radio Frequency (RF) transceiver. The UWB is based on Decawave’s DWM1000 transceiver,

which provides accurate distance measurements with a high power consumption. To reduce the power con-

sumption, the data transmission between nodes is performed by the low-power RF transceiver. Using this

architecture, the experimental validation shows a good performance in both power consumption and accuracy.

The system’s overall power consumption is reduced as well as a 10-centimeter accuracy level is achieved.

1 INTRODUCTION

Internet of Things (IoT) is nowadays of high inter-

est in several sectors, including smart home, precision

agriculture and e-health and industry 4.0 (Kanoun

et al., 2021). In particular, Wireless Sensor Networks

(WSNs) are gaining importance as enabling technol-

ogy for monitoring and decision-making (Khriji et al.,

2018), (El Houssaini et al., 2018a). With the ad-

vancements in WSNs, indoor localization becomes

inevitable to track objects in indoor environments by

offering many services, including predictive mainte-

nance, store tracking and state machine monitoring

oﬂinger et al., 2018), (Obeidat et al., 2021) and

(El Houssaini et al., 2018b). Therefore, different posi-

tioning techniques are investigated to locate wireless

sensor nodes.

Determining the current location of sensor nodes

with high accuracy is the principal target of most ex-

isting localization studies. However, this requires a

high power demand. Due to the energy constraint

of sensor nodes, the energy consumption needs to

https://orcid.org/0000-0002-0562-0116

https://orcid.org/0000-0003-1764-6601

https://orcid.org/0000-0002-7166-1266

be highly considered while performing the localiza-

tion process. Only a few research (Kaur and Mal-

hotra, 2016), (Cheng, 2021), and (Amutha et al.,

2020) address the energy aspects of localization along

with the accuracy. Therefore, the main aim of this

paper is to establish a trade-off between accuracy

and energy consumption for indoor positioning. In

this context, an energy-efﬁcient accurate positioning

system is designed, which embeds the Ultra Wide

Band (UWB) compliant wireless transceiver mod-

ule, DWM1000 and the low-power wireless mod-

ule, panStamp. The ﬁrst module determines the dis-

tance between nodes with high accuracy. The distance

calculation as well as performing different packets

transmission/reception are carried out with the low-

power radio transceiver in-built on the panStamp

node. Thus, it reduces the overall power consump-

tion needed for node positioning. To conclude, the

key contributions of this work are:

• A panoramic view of indoor positioning technolo-

gies

• Design of energy-efﬁcient accurate positioning

system involving UWB wireless module and low-

power RF transceiver

The rest of the paper is organized as follows. Section

Khriji, S., El Houssaini, D. and Kanoun, O.

Design of an Energy-efﬁcient Hybrid UWB-RF Indoor Distance Estimation System.

DOI: 10.5220/0011012700003118

In Proceedings of the 11th International Conference on Sensor Networks (SENSORNETS 2022), pages 249-255

ISBN: 978-989-758-551-7; ISSN: 2184-4380

249

2 presents an overview of indoor positioning tech-

nologies with highlighting the use of the UWB tech-

nology. The design and implementation of the pro-

posed distance measurement architecture is provided

in section 3. Section 4 presents the experimental re-

sults of the designed system in terms of accuracy and

power consumption. Section 5 concludes the paper.

2 OVERVIEW OF INDOOR

POSITIONING

TECHNOLOGIES

Indoor positioning can be divided into building de-

pendent and building independent (Alariﬁ et al.,

2016) technologies (see Fig. 1). Building dependent

technologies rely on the building, where they are go-

ing to be installed. They are based on an available

technology in the building, or on the building’s struc-

ture and layout. This category can be further clas-

siﬁed into two classes: Dedicated infrastructure re-

quired and building’s infrastructure utilized. The in-

door positioning that needs a dedicated infrastructure

can be RFID, infrared, ultrasonic, Zigbee and UWB.

Indoor positioning technologies

Building dependent Building independent

Dedicated

infrastructure required

Building’s

infrastructure utilized

RFID

UWB

Infrared

Ultrasonic

Zigbee

WiFi

Bluetooth

Cellular-based

Dead reckoning

Image-based technology

Figure 1: Indoor positioning technologies.

RFID systems rely on proximity detection, and

hence have an accuracy of dm-m level (Xu et al.,

2017). A larger area requires several RFID tags, re-

sulting on increasing the cost as well as power con-

sumption. UWB systems can work in Non-Line of

Sight (NLOS) conditions. UWB operates within 3.5

to 6.5 GHz, which does not interfere with other fre-

quencies. The Infrared (IR) wireless communication

uses the invisible spectrum of light (Arbula and Lju-

bic, 2020). Differential phase-shift and angle of ar-

rival (AoA) localization methods are commonly used

with IR technology. Another technology requiring the

building’s infrastructure is the ultrasonic technology.

It is based on the ToF (Time of Flight) localization

method. Therefore, ultrasound impulses are carried

out by the tag to measure the distance to the anchor.

This technology has a short-range, and does not inter-

fere with electromagnetic waves.

The Zigbee technology can be also used to de-

termine the nodes’ positions. This technology is de-

signed for applications requiring low throughput and

low power consumption. It comprises a microcon-

troller and a multi-channel bidirectional radio on a

single piece of silicon (Aykac¸ et al., 2017). Typically,

Received Signal Strength (RSS) values have been em-

ployed to determine the distance between Zigbee-

based sensor nodes.

WiFi, Bluetooth and cellular-based technologies

refer to the indoor positioning technologies that use

the building’ infrastructure. WiFi positioning systems

can reach up to a range of 50 m, but the accuracy ob-

tained is very low (in meters).

Furthermore, building independent technologies

can be a dead reckoning and image-based technolo-

gies. In dead reckoning, the current position of an

object can be determined by its previous position, di-

rection of travel and speed (Zhou et al., 2015). On the

other hand, image-based technologies are often sup-

ported by cameras and computer vision technologies

(Wu et al., 2018). A summary of different indoor po-

sitioning technologies is depicted in Table. 1.

Table 1: Indoor positioning technologies.

Technology Range Accuracy Measuring Principle

RFID 50 m dm-m Cell of Origin

Ultra-Wide Band 50 m cm TWR,ToA,TDoA

Infrared 5 m cm-m Differential phase-shift, AoA

Ultrasound 10 m cm ToA, TDoA

Zigbee 20 -30 m cm Trilateration, RSSI

WiFi 50 m m RSSI,ToA,TDoA

Bluetooth 10 - 100 m m Proximity, RSS

Dead reckoning - m Tracking

Cameras 10 m mm-dm Pattern recognition

3 DESIGNED HYBRID SYSTEM

The main aim of the proposed system is to determine

the distance between sensor nodes with a high accu-

racy and less power consumption. The block diagram

of the localization system architecture is depicted in

Fig. 2.

SPI

µC

UWB

Transceiver

UWB

Transceiver

µC

SPI

UWB

3.5 - 6.5 GHz

868 MHz

Tag

Anchor (x3+)

Figure 2: UWB localisation system architecture.

Therefore, the UWB system consists of a tag and

EWSN-IoT 2022 - Special Session on Energy-Aware Wireless Sensor Networks for IoT

250

several anchors. Both tag and anchor consist of the

same hardware, with a microcontroller and an UWB

transceiver. The UWB technology is used to pro-

vide a high accuracy with limited range. A low-

power microcontroller is connected to the UWB via

Serial Peripheral Interface (SPI) to ensure the compu-

tation of distance measurements. A low-power radio

transceiver with a frequency of 868 MHz is used to

exchange other data needed by the application.

DWM1000

Panstamp

NRG 2.0

IRQ

SCK

MISO

MOSI

CSn

D22

RST

D21

Figure 3: Interfacing DWM1000 with Panstamp NRG 2.0.

To determine the distance between the tag

and anchor nodes, the wireless transceiver module

DWM1000 is used, which is an IEEE802.15.4-2011

UWB compliant module. This transceiver is used

only for communicating the distance between two

nodes (DWM, 2021). The main advantage of the se-

lected UWB transceiver is the high data rate commu-

nication, as it can transmit pulses with a bandwidth

up to 6.8 Mbps. It determines also the positions of

objects with an accuracy of 10 centimetres in NLoS

conditions. It supports TDoA and ToF location tech-

niques. This transceiver consists of the Two-Way

Ranging (TWR) method, which determines the ToA

of the UWB RF signal. Then, it calculates the dis-

tance between nodes by multiplying the time by the

speed of light.

To measure the distance between two nodes, three

messages need to be exchanged (see Fig. 4). The tag

sends a Poll message to know the address of the an-

chor in time referred to Time of Sending Poll (T

The time of the Poll reception (T

) is recorded by

the anchor, which replies with the response message

at time T

. By receiving the response message, the

tag node records the time T

, and composes a Final

message, including its address, T

, T

and T

infor-

mation. Based on the timestamps, the communicated

signals between the tag and anchor are used to mea-

sure the ToA of the UWB signal, as presented in equa-

tion 1. Thereby, the distance is determined by multi-

plying the ToA of the UWB signal with the speed of

light c as depicted in equation 2.

Tag

Anchor

Poll

ToA

Response

Final

ToA

Report distance

Figure 4: Distance measurement using TWR technique im-

plemented on DWM1000 transceiver.

ToA =

−T

)−(T

−T

)+(T

−T

)−(T

−T

)

(1)

distance = c × ToA (2)

Due to limited ressources of sensor node, the

power consumption is a key feature in wireless sen-

sor networks. Thereby, choosing low-power hardware

can reduce the overall node power consumption. In

this direction, the low-power RF-based wireless sen-

sor node, panStamp, is used. The module is based

on the CC430F5137 System on chip (SoC), including

the MSP430 microcontroller and CC1101 radio chip

(Salem et al., 2019), which operates at different fre-

quencies like 433 MHz, 868 MHz and 915 MHz. It

supports SPI communication, which enables the com-

munication with the DWM1000 module. The con-

nections are shown in Fig. 3. The panStamp com-

municates with the DWM1000 at 4 MHz SPI fre-

quency. Apart from SPI connection, a reset line (RST)

and interrupt request (IRQ) lines are connected to the

DWM1000. The RST line can be pulled low to re-

set the DWM1000 module. The IRQ is an input to

the Panstamp and is active high. An interrupt is gen-

erated by the DWM1000 to notify Panstamp at each

event such as transmission, reception or wake up.

Fig. 5 shows the block diagram of the tag node.

The initialization step includes the conﬁguration of

the SPI interface with the microcontroller. The SPI

communication is initiated at the desired rate. During

this phase, the DWM1000 transmission rate and fre-

quency are conﬁgured. A TimerDelay is set, which

deﬁnes the rate of the distance measurement. Every

time, the internal clock of the microcontroller reaches

the TimerDelay, a ranging cycle is initiated. This

Design of an Energy-efﬁcient Hybrid UWB-RF Indoor Distance Estimation System

251

starts with a polling message sent by the tag, and

ends after receiving the range report from the anchor.

The tag is responsible for sending timestamps in the

range message. After each range measurement, the

tag switches to the sleep mode and wakes up again

before initiating the measurement sequence.

Initialization

TimerDelay

elapsed

Counter for

Blink

reached

Transmit Blink

Transmit Poll

PollAck

Received

Wait

Transmit Range

RangeReport

Received

Wait

Display Range

yes

Figure 5: Ranging algorithm for the tag node.

As depicted in Fig. 6, the anchor node starts in the

receiving mode, waiting for a Poll message from the

tag. On receiving this message, it acknowledges by

transmitting a PollAck message and waits for a range

message. The range message includes all the times-

tamps calculated on the tag side and, along with the

timestamps measured at the anchor, a Time-of-Flight

(ToF) is computed. This ToF is converted into the dis-

tance, and thus transmitted back to the tag as a range

report message.

4 EXPERIMENTAL RESULTS

The designed indoor localization system is evaluated

in terms od power consumption and accuracy.

Init

Poll

Received

Transmit PollAck

Range

Received

Compute Range,

Transmit RangeReport

yes

Figure 6: Ranging algorithm for the anchor node.

4.1 Power Consumption

The main aim of using a panStamp as a microcon-

troller for DWM1000 is its low-power consumption

along with its wireless capabilities. The Rx and Tx

current for panStamp is 18 mA and 36 mA respec-

tively, which is far less than the DWM1000 current

consumption (140 mA and 160 mA). Thus, messages

that do not play a part in the range calculation can be

transmitted using the radio module in panstamp in-

stead of DWM1000. Messages that can be replaced

are Blink, RangingInit, RangeReport.

Since receiving a message consumes a lot of

power, we turn on the receiver exactly at the time a

message is expected. Thus, scheduling of the mes-

sages is very important. Also, when in sleep mode,

the microcontrollers of tag and anchors need to com-

municate with each others to wake up the DWM1000

at the right time. In the case of panStamp, this

is carried out using the radio module available on

panStamp. The maximum current consumption as per

the datasheets of DWM1000 and panStamp is given

in Table 2. Based on the timestamps generated from

Panstamp using the serial monitor, the total current

consumption of the system can be estimated accord-

ingly as the maximum current consumption is known.

For this, consider a measurement cycle between a Tag

and Anchor. As the radio signals travel at the speed

of light, the signal transmission time in the air is very

low and would not differ much for the operating range

of 50 meters. Various events are recorded using the

serial monitor of the Panstamp and a summary of time

and current required for each event in tag and anchor

EWSN-IoT 2022 - Special Session on Energy-Aware Wireless Sensor Networks for IoT

252

can be seen in Fig. 3, Fig. 8 and Tables 3 and 4.

Table 2: Maximum current consumption for DWM1000 and

panStamp.

Device Mode DWM1000 Panstamp

Tx 140 mA 18 mA

Rx 160 mA 36 mA

Deepsleep 200 nA 2 uA

Idle 13.4 mA 2 mA

Wakeup Tx

DW1000 Wakeup

DW1000 Init

Tx Poll

Rx wait

Rx PollAck

Tx wait

Tx Range

Rx wait2

Rx Report

Idle

100

120

140

160

4 ms

3 ms 46 ms

1 ms

5 ms

1 ms

7 ms

1 ms 12 ms 1 ms

919ms

15.4 15.4

142

15.4

162

15.4

142

2 2 2 2 2 2 2 2

Event

Current (mA)

Panstamp

DWM1000

Figure 7: Interfacing DWM1000 with Panstamp NRG 2.0.

Wakeup Rx

DW1000 Wakeup

DW1000 Init

Rx Poll

Tx wait

Tx PollAck

Rx wait

Rx Range

Compute Range

Tx Report

Idle

100

120

140

160

8 ms 3 ms 46 ms

1 ms

6 ms

1 ms

6 ms

1 ms

8 ms 6 ms

914ms

15.4 15.4

162

15.4

142

15.4

162

15.4

2 2 2 2 2 2 2 2

Event

Current (mA)

Panstamp

DWM1000

Figure 8: Estimated current consumption in anchor for one

cycle.

Each cycle includes a range measurement between

a Tag and Anchor using DWM1000 and panStamp.

Initially, both DWM1000 modules are in deep sleep

mode and the panStamp node of the tag is in Idle

mode, while the panStamp node of the anchor is in re-

ceiving mode. Therefore, the total current consump-

tion is 2 mA for the tag, and 36 mA for the anchor

node. To initiate a measurement, the tag sends a

wake-up signal to the anchor via the radio module of

panStamp. This message includes 2 bytes signal with

the address of the anchor node and message ID for

wake-up command. Immediately, after the transmis-

sion from the tag, the radio module of panStamp is

turned off and an SPI signal from panStamp wakes up

the DWM1000. This takes approximately 3 ms. As

the DWM1000 is in deep sleep mode, it needs to be

reconﬁgured. The initialization of the module takes

around 46 ms. The DWM1000 starts in idle mode

and consumes around 13.4 mA current.

Table 3: Estimated current consumption of the tag for one

cycle.

Event

panStamp DWM1000 Time

(ms)

Total

Current

(mA)

State

Current

(mA)

State

Current

(mA)

Wakeup Tx Tx 18 Deepsleep 0.0002 4 18.0002

DW1000 wakeup Idle 2 Idle 13.4 3 15.4

DW1000 init Idle 2 Idle 13.4 46 15.4

Tx Poll Idle 2 Tx 140 1 142

Rx wait Idle 2 Idle 13.4 5 15.4

Rx PollAck Idle 2 Rx 160 1 162

Tx wait Idle 2 Idle 13.4 7 15.4

Tx range Idle 2 Tx 140 1 142

PS Rx wait Idle 2 Deepsleep 0.0002 12 2.0002

Rx RangeReport Rx 36 Deepsleep 0.0002 1 36.0002

Idle Idle 2 Deepsleep 0.0002 919 2.0002

On receiving the wake-up command from the ra-

dio module of panStamp, the anchor node wakes up

the DWM1000 module via the SPI communication.

The receiver of the panStamp radio is also turned off

to save power. After completing the initialization of

the DWM1000, the receiver is turned on, and the tag

transmits the ﬁrst message from DWM1000, the Poll

message. Thus, the anchor receives this message at

the same time. The time difference between turning

on the receiver and actual reception is calculated at

the anchor side, which is approximately 1 ms. There-

fore, the 160 mA consumption of the DWM1000 in

the receiving state only occurred for 1 ms. The trans-

mission of 128 symbol preamble and 20 bytes of data

requires only 0.2 ms when operating at 6.8 Mbps data

rate. This means that the transmitter of the DWM1000

would be operating only for 0.2 ms consuming 140

mA current. All the transmissions in the DWM1000

are carried out at this rate, but for calculating the cur-

rent consumption, we assume the transmission time

as 1 ms for all the messages.

The DWM1000 is a half-duplex transceiver that

can either transmit or receive a message at a single

time. By default, the module switches to its idle state

after transmitting or receiving a message. After re-

ceiving the Poll message, the anchor saves the times-

tamp in the panStamp memory, and starts transmitting

the PollAck message. This time is termed as Tx wait,

where the DWM1000 is in idle state and ready for

Design of an Energy-efﬁcient Hybrid UWB-RF Indoor Distance Estimation System

253

Table 4: Estimated current consumption of the anchor for

one cycle.

Event

Panstamp DWM1000 Time

(ms)

Total

Current

(mA)

State

Current

(mA)

State

Current

(mA)

Wakeup on Rx Rx 36 Deepsleep 0.0002 8 36.0002

DW1000 wakeup Idle 2 Idle 13.4 3 15.4

DW1000 init Idle 2 Idle 13.4 46 15.4

Rx Poll Idle 2 Rx 160 1 162

Tx wait Idle 2 Idle 13.4 6 15.4

Tx PollAck Idle 2 Tx 140 1 142

Rx wait Idle 2 Idle 13.4 6 15.4

Rx range Idle 2 Rx 160 1 162

Compute range Idle 2 Idle 13.4 8 15.4

Tx RangeReport Tx 18 Deepsleep 0.0002 6 18.0002

Idle Idle 2 Deepsleep 0.0002 914 2.0002

its next transmission. The tag also saves the times-

tamp it sent the Poll message and is now waiting for

its next reception. Since the PollAck is not sent im-

mediately, the tag stays in idle state to save power,

which is termed as Rx wait. After waiting for a few

seconds, the tag switches to receiving mode to receive

the PollAck message. Higher Rx wait time refers to

a lower reception time for the tag resulting on power

saving. The Rx wait time is gradually increased in

the tag until the message reception is possible. Once

stable receptions are available, the time is ﬁxed as Rx

wait time.

Upon receiving the PollAck, the tag records this

timestamp. This timestamps has to be transmitted to

the anchor. During the Tx wait, the tag gathers all the

available timestamps and prepares the ranging mes-

sage. This took around 7 ms for the tag. Once the

message is ready, the tag transmits this range message

and sets the DWM1000 to deep sleep. The Rx wait

time for this message was also optimized on the an-

chor side, where the DWM1000 was idle for 6 ms. On

receiving the range message, the anchor has to calcu-

late the exact range between the tag and anchor based

on the timestamps available and those received from

the range message. This calculation took around 8 ms

after which the DWWM1000 is switched from idle

state to deep sleep mode. Since the range is already

calculated, both the DWM1000 are not required to be

active and are sent to deep sleep. At this point, only

the anchor knows the range between the two points,

and has to transmit this information to the tag. The

transmission of the RangeReport happens over the

panStamp radio module. To receive this message, the

panStamp receiver at the tag has to be turned on as

well. The Rx wait time is calculated and upon re-

ceiving the RangeReport message, the Panstamp Ra-

dio module is turned off. At this point, both the tag

and anchor have calculated the range, and enter the

idle state, waiting for the next measurement cycle.

The next measurement cycle starts with the wake-

up signal from the tag, but to receive this message

the panStamp radio has to be in the receiving mode.

Therefore, depending upon the update rate required,

the receiver at the anchor is automatically turned on

after a ﬁxed delay time. The panStamp can also be

used in Wake-on-Radio (WOR) mode to continuously

check for the reception after certain time intervals.

By combining the total estimated current con-

sumption with the time required for each state, an

average current consumption can be calculated. For

ranging a new measurement every second, the aver-

age current consumption estimated is 3.35 mA for tag

and 3.75 mA for anchor. This is a considerable reduc-

tion in the power consumption for DWM1000 mod-

ules along with the microcontroller. Reducing the up-

date rate would further reduce power consumption.

4.2 Distance Measurements

Fig. 9 shows the distance measurement using

DWM1000 interfaced with a panStamp module.

0 20 40

80 100 120 140

160

180 200 220 240

260

280 300

100

150

200

250

300

Distance (cm)

Actual distance

Measured distance

Filtred distance

Figure 9: Measured distance using DWM1000 with

Panstamp.

It can be seen that the raw measured distance is

far away from the actual distance, but follows a lin-

ear pattern. This can be corrected by adjusting the

antenna delay value provided in the timestamp cal-

culation. The antenna itself has a ﬁxed transmission

delay. To compensate for this delay in DWM1000, an

integer value is set during the initialization. Since the

plot is linear for the raw measured data, a linear ﬁlter

is applied during the calculation of the distance to get

more accurate readings. Using least squares ﬁtting,

the following polynomial is obtained for the curve:

Filter dist = 0.53375 ×Measur dist +20.469 (3)

The raw measurements are then linearised using this

ﬁlter. It can be seen from Fig. 9 that an accuracy of

10 cm can be achieved in indoor environment using a

DWM1000 with panStamp.

EWSN-IoT 2022 - Special Session on Energy-Aware Wireless Sensor Networks for IoT

254

5 CONCLUSION

An Ultra-wideband localization system is designed

using Decawave’s DWM1000 and panStamp wire-

less node. The panStamp enables a second channel

for the communication within the nodes, thus keep-

ing the UWB messages to the minimum. Two fre-

quency bands are utilized by the system, 6.5 GHz

for DWM1000 and 868 MHz for the panStamp radio.

The Panstamp operates at very low current, which re-

duces the overall current consumption of the system.

It is estimated that the system consumes a total of 3.35

mA current for tag and 3.75 mA current for anchor

during operation at 1 Hz update frequency. Even with

the low power microcontroller, a good performance

of the system is achieved. Tests performed indoors

within a distance of 3 meters show an accuracy of 10

cm.

REFERENCES

(2021). DW1000 Datasheet. decaWave.

Alariﬁ, A., Al-Salman, A., Alsaleh, M., Alnafessah, A.,

Al-Hadhrami, S., Al-Ammar, M. A., and Al-Khalifa,

H. S. (2016). Ultra wideband indoor positioning tech-

nologies: Analysis and recent advances. Sensors,

16(5):707.

Amutha, J., Sharma, S., and Nagar, J. (2020). Wsn strate-

gies based on sensors, deployment, sensing models,

coverage and energy efﬁciency: Review, approaches

and open issues. Wireless Personal Communications,

111(2):1089–1115.

Arbula, D. and Ljubic, S. (2020). Indoor localization based

on infrared angle of arrival sensor network. Sensors,

20(21):6278.

Aykac¸, M., Erc¸elebi, E., and Aldin, N. B. (2017). Zigbee-

based indoor localization system with the personal dy-

namic positioning method and modiﬁed particle ﬁl-

ter estimation. Analog Integrated Circuits and Signal

Processing, 92(2):263–279.

Cheng, Q. (2021). A robust indoor localization algo-

rithm for wsn in los and nlos environment. In

2021 IEEE 11th International Conference on Elec-

tronics Information and Emergency Communication

(ICEIEC) 2021 IEEE 11th International Conference

on Electronics Information and Emergency Commu-

nication (ICEIEC), pages 91–94. IEEE.

El Houssaini, D., Khriji, S., Besbes, K., and Kanoun, O.

(2018a). Wireless sensor networks in agricultural ap-

plications. In Energy Harvesting for Wireless Sensor

Networks, pages 323–342. De Gruyter Oldenbourg.

El Houssaini, D., Mohamed, Z., Khriji, S., Besbes, K.,

and Kanoun, O. (2018b). A ﬁltered rssi model based

on hardware characteristic for localization algorithm

in wireless sensor networks. In 2018 32nd Interna-

tional Conference on Advanced Information Network-

ing and Applications Workshops (WAINA), pages 118–

123. IEEE.

oﬂinger, F., Bordoy, J., Zhang, R., Bannoura, A., Si-

mon, N., Reindl, L. M., and Schindelhauer, C. (2018).

Localization system based on ultra low-power radio

landmarks. In SENSORNETS, pages 51–59.

Kanoun, O., Khriji, S., Naifar, S., Bradai, S., Bouattour, G.,

Bouhamed, A., El Houssaini, D., and Viehweger, C.

(2021). Prospects of wireless energy-aware sensors

for smart factories in the industry 4.0 era. Electronics,

10(23):2929.

Kaur, R. and Malhotra, J. (2016). Efﬁcient localization in

wsn based on structured deployment of anchor nodes

in dv-hop. International Journal of Future Generation

Communication and Networking, 9(3):155–166.

Khriji, S., El Houssaini, D., Kammoun, I., and Kanoun, O.

(2018). A fuzzy based energy aware unequal clus-

tering for wireless sensor networks. In International

Conference on Ad-Hoc Networks and Wireless, pages

126–131. Springer.

Obeidat, H., Shuaieb, W., Obeidat, O., and Abd-Alhameed,

R. (2021). A review of indoor localization techniques

and wireless technologies. Wireless Personal Commu-

nications, pages 1–39.

Salem, J. B., Khriji, S., Baklouti, M., Kammoun, I., and

Kanoun, O. (2019). Testbed implementation of a

fuzzy based energy efﬁcient clustering algorithm for

wireless sensor networks. In 2019 16th Interna-

tional Multi-Conference on Systems, Signals & De-

vices (SSD), pages 351–356. IEEE.

Wu, Y., Tang, F., and Li, H. (2018). Image-based camera

localization: an overview. Visual Computing for In-

dustry, Biomedicine, and Art, 1(1):1–13.

Xu, H., Ding, Y., Li, P., Wang, R., and Li, Y. (2017). An rﬁd

indoor positioning algorithm based on bayesian prob-

ability and k-nearest neighbor. Sensors, 17(8):1806.

Zhou, Z., Chen, T., and Xu, L. (2015). An improved

dead reckoning algorithm for indoor positioning based

on inertial sensors. In International Conference of

Electrical, Automation and Mechanical Engineering

(EAME 2015), pages 369–371.

Design of an Energy-efﬁcient Hybrid UWB-RF Indoor Distance Estimation System

255