A NOVEL ADAPTIVE - WAVELET BASED DETECTION ALGORITHM FOR CHIPLESS RFID SYSTEM

Dhinaharan Nagamalai et al. (Eds) : CoSIT, SIGL, AIAPP, CYBI, CRIS, SEC, DMA - 2017
pp. 29– 38, 2017. © CS & IT-CSCP 2017 DOI : 10.5121/csit.2017.70404
A NOVEL ADAPTIVE - WAVELET BASED
DETECTION ALGORITHM FOR CHIPLESS
RFID SYSTEM
Meriam A. Bibile and Nemai C. Karmakar
Department of Electrical and Computer Systems Engineering,
Monash University, Melbourne, Australia
ABSTRACT
In this paper, a novel wavelet-based detection algorithm is introduced for the detection of
chipless RFID tags. The chipless RFID tag has a frequency signature which is identical to itself.
Here a vector network analyser is used where the received backscatter signal is analysed in
frequency domain. Thus the frequency signature is decoded by comparing the wavelet
coefficients which identifies the bits accurately. Further, the detection algorithm has been
applied for the tag detection under different dynamic environments to check the robustness of
the detection algorithm. The new method doesn’t rely on calibration tags and shows robust
detection under different environments and movement.
KEYWORDS
Chipless RFID, wavelet, backscatter signal, frequency domain
1. INTRODUCTION
The chipless RFID reader extracts the backscattered signal and decodes the tag ID. This is an
ongoing challenge, as the detection procedure for a chipless RFID tag has more complexities
compared to a conventional RFID tag. The signal collides with other scatterers or tags which give
a ‘clutter' signal with interference. A number of detection techniques have been applied to
achieve an accurate result of its tag ID.
The basic detection technique is based on comparing the received data with threshold values
obtained by calibration. It is, therefore, a basic approach and it does not possess the flexibility
and adaptability required in the detection process to address errors due to a dynamic environment
[1]. Different types of detection algorithms and decoding techniques have been revealed in the
past few years.
Moving average technique is a simple de-noising technique which removes noises by acting as a
low pass filter. An 11 sample averaging moving average filtering has been successfully
implemented on a low-cost mid-range microcontroller having low processing power capabilities,
and a smoothened waveform is resulted after using this filtering technique. Hilbert transform

30 Computer Science & Information Technology (CS & IT)
(HT) is a complex analytical signal processing technique [2]. This technique has been used to
reconstruct the frequency signatures of the chipless tags. It has been experimentally proven that
HT provides the extraction of the amplitude and phase functions of the frequency signature.
The signal space representation of chipless RFID tags uses an efficient mathematical model to
decode information in a chipless RFID tag [3-4]. The frequency signatures are represented by a
matrix which is composed of orthonormal column vectors and a singular value matrix. The
constellation of signal points are plotted with a basis function. It can be seen that as the number
of bits increase this method will face limitations. Matrix pencil method (MPM) and Short time
matrix principle method (STMPM) are two more detection techniques that have been applied for
chipless RFID systems [5-6]. These two techniques are applied in the time domain and are
mentioned as accurate detection techniques in extracting the carrier to noise ratio (CNR) of the
response. Detection is performed by extracting the poles and residues from the backscattered
signal using the Matrix Pencil Algorithm. A Maximum Likelihood (ML) based tag detection
technique and Trellis decoding technique has been developed where detection error rate is
compared with the bit to bit detection [7]. It has been found that ML detection has the best
performance. It also reports that the computational complexity is higher in ML detection
technique than Trellis detection technique.
The main aim of this paper is to develop a detection algorithm which can be practically applied in
firmware. As many of the above algorithms are highly complexed in implementing in the chipless
RFID reader. Also, most of them have the limitation in the number of bits that can be detected. In
this paper, we have developed a novel wavelet which suits the chipless RFID received signal to
detect the frequency ID of the chipless RFID tag.
2. SYSTEM MODEL AND DESIGN
In this section the system model of the chipless RFID system used in this analysis is discussed. A
description of the backscattered signal is also given and the backscattered signal of the used tag is
analysed using simulation results obtained through CST microwave studio.
2.1. Experimental Model
Figure 1 –Chipless RFID system

Computer Science & Information Technology (CS & IT) 31
The experiment is performed using a 5-bit spiral resonator design. Each resonator has its own
resonance frequency and has a diameter of 5mm. The tag with 5 bits has a dimension of 2cm ×
2cm.The tag design is not presented in this paper to due to the confidentiality of the tag design.
The tag RCS is shown in Fig 2. A patch antenna is used for the measurements, and the reading is
taken by loading the antenna with a tag using vector network analyzer (VNA) as shown in Fig 1.
The tag is loaded above the patch antenna and is placed on a piece of foam with 1cm thickness.
Backscattered signals from chipless RFID tags are very weak, and the detection process at the
RFID reader is extremely susceptible to noise. This is because the information is represented as
analog variations as opposed to a modulated digital stream of data as in conventional wireless
communication of information where error correction schemes can be used to detect the presence
of errors and discard erroneous data packets.
According to Fig 2, the resonant frequencies of the tag are located at 4.15, 4.70, 5.37, 6.10 and
6.95GHz respectively. The other local maxima detected in the signal are due to the amplitude
spectrum of the background and antenna S11. These spurious peaks need to be carefully filtered
to detect the correct frequency signature of the chipless RFID tag.
Figure 2 – S11 Magnitude (dB) vs. frequency of the 5-bit resonant tag
2.2. Wavelet Design
In this paper, a novel wavelet is adopted to detect the peaks of the backscattered signal. It is based
on the Gaussian function which is given by (1).
݂ሺ‫ݔ‬ሻ = ܽ × ݁
షሺೣష್ሻమ
మ೎మ
(1)
where a is the height, b is the position and c is the width of the wavelet. The values a, b and c are
adjusted to maximize the detection.
The novel wavelet design is shown in Fig 3. The width of the wavelet is adaptive according to the
number of bits in the tag. In Fig 3, the width of the wavelet has been taken as 250MHz giving a

wider bandwidth to the wavelet. Further for the detection of the results this has been changed to
100MHz giving better resolution of the detected bits. This is an advantage of this algorithm as
higher number of bits can be detected with better resolution.
The detected signal is compared with the wavelet, and a coefficient is defined to represent how
closely correlated the wavelet is with each part of the signal. The larger the coefficient is in
absolute value, the more the similarity appears. The wavelet is shifted until it covers the whole
signal. The threshold coefficient is defined after performing the experiment for the tag with bit
‘11111’ number of times under different environmental changes. By the variation in the detected
signal under different circumstances a minimum and maximum coefficient is set giving rise to a
confident band for detection.
Figure 3 – New adopted wavelet pattern using Gaussian function
2.3. Flowchart for Detection Algorithm
The flowchart given in Fig 4, shows the steps of the developed algorithm and how it is applied to
the received signal.
First the measured results using VNA are loaded into Matlab. The post processing is performed
using Matlab programming software. The program can be directly downloaded to the
microcontroller of the frequency domain reader developed by the research group which will be
implemented in the next step of this research.
A baseline removal is performed on the first results to remove the nulls of the antenna S11. Then
the wavelet is defined and is integrated with the resulting signal. The tag ID is decoded based on
the condition band defined for the detection.
Experiments have been performed under different indoor dynamic environments such as placing
clutter objects above the chipless RFID tag. Fig 8, shows the placing of hand 7cm above
covering the tag adding back reflection to the received signal. Similarly, a copper plate was also
placed above the tag at 7cm, and the received signal was observed using the VNA.

Computer Science & Information Technology (CS & IT)
Figure 7 –
Figure 8 – placing hand on top of the tag at 7cm distance giving attenuation to the signal received at the
Flowchart for applying the detection algorithm
placing hand on top of the tag at 7cm distance giving attenuation to the signal received at the
VNA
33
placing hand on top of the tag at 7cm distance giving attenuation to the signal received at the

3. RESULTS
The measured results of the tag detection are shown in Fig 9.
Figure 9 – Measured results (a) of the original tag at 1.5cm above the patch antenna, (b) with hand placed
above the tag at a distance of 7cm, (c) with copper plate placed on top of the tag at a distance of 7cm, (d)
by placing the tag at the corner of the patch antenna
It shows the received signals from the reader before applying the algorithm. In the original signal
with the tag, we can identify the 5-bit resonances from the received signal itself. But when the
hand, the copper plate is placed on top of the tag we can see that the received signal has increased
in magnitude and has been shifted up. Also, peaks are not distinguishable at a glance. When the
tag is placed in the corner of the antenna which implies the tag is moved out of the beam width of
the antenna, we can again see a distortion in the signal as well as shifting the signal down in
magnitude.
4. ANALYSIS
The detected signal after applying the detection algorithm is analysed in this section. The data is
then interpreted to the end user. The results obtained after applying the developed wavelet based
detection algorithm is given in Fig 10.

Figure 10 – Wavelet coefficients of the detected signals after applying the developed algorithm.
(a) Original tag detection (b) Hand on top of tag (c) Copper plate on top of tag (d) Tag moved to the corner
of antenna
The coefficients obtained are analysed in Table 1. The coefficients of the original tag are set to 10
which is taken as the threshold value of detection for the original tag with bits ‘11111’. A
condition band is set for the maximum and minimum detectable coefficient bands. For the
chipless RFID tag used in this experiment, the condition band is set between ±8, which gives +18
as the maximum value of detection as bit ‘1' and +2 as the minimum level of detection as bit ‘1'.
If the coefficient is below or above these values, it will detect its ID as a bit ‘0'.
When the hand is placed, we can see that the coefficients are varying and the maximum
coefficient reaches 17.35. When the copper plate is placed, it reaches a maximum value of 10.32
and the tag moved to the corner of the antenna gives a coefficient of 17.35. According to these
results, the frequency ID of the tag is detectable as ‘11111' under all the applied conditions.

Table 1 – Wavelet coefficients for (a) Original tag detection – when the antenna is loaded with tag only (b)
Hand on top of tag – when the hand is placed at 5cm distance above the tag (c) Copper plate on top of tag -
when a copper plate of dimensions 10× 10× 0.2mm is placed at 5cm distance above the tag (d) Tag moved
to the corner of antenna - when the tag is placed in the corner of the antenna
Resonance
Freq.
(GHz)
Original tag
detection
(Coefficient value)
Hand on top
of tag
Copper plate on
top of tag
Tag moved to
corner of
antenna
4.15 10.14 14.53 10.15 14.53
4.70 10.25 17.35 10.25 17.35
5.37 10.32 12.88 10.32 12.88
6.10 10.29 13.26 10.29 13.26
6.95 10.27 16.72 10.27 16.72
The frequency domain reader developed in MMARS lab uses the same patch antenna design for
detection and is working in the frequency range of 4 – 7.5GHz. Therefore future direction of this
research will be the implementation of the novel detection algorithm in firmware.
5. CONCLUSIONS
According to these results, it can be seen that a hand or conductive material (Cu) placed in the
close vicinity of the tag (less than 10cm) will have an effect on the magnitude of the tag
measurement. Still applying the novel detection algorithm the tag can be identified even if the
conductive material is placed covering the tag. Also, when the tag is moved around the surface
area covered by the antenna, the tag can be detected using this detection algorithm. Since the
experiment was conducted by placing the clutter only on one side of the tag, more experiments
should be performed by varying the position, size, material, etc. to conclude the probability of
error due to the different material surrounding the tag.
An advantage of this detection algorithm is that it could detect up to a minimum of 1dB
magnitude variation with added noise. The adaptability of the wavelet is another advantage of this
algorithm as it can be applied for any type of tag in time domain or frequency domain and could
cover the required frequency range. Also the detected bit resolution is high, which leaves room
for the applications with higher number of bits in future.
When compared with other algorithms this has the most adaptability to any kind of chipless RFID
system. Comparing the number of bits of a tag, the presented algorithm can be applied for higher
number of bits whereas algorithms developed by Kalansuriya[3], Rezaiesarlak [6] and
Divarathna[7] are implemented for low number of bits due to their mathematical complexity. As
the width of the wavelet used in the detection is only 10MHz, with a guard band of another
10MHz a 20bit tag can be easily detected in the given frequency range of 4-8GHz.

The algorithm can be adapted into a wider bandwidth of frequencies and can be applied for any
other chipless tag design in time or frequency domain.
Most of the developed detection algorithms in this research area are tested only for simulation
results or for measured results using a vector network analyser. Presently the algorithm has been
programmed into the microcontroller of the chipless RFID reader.Therefore this tag detection
algorithm can be used further to investigate the robustness of the chipless RFID reader system
under a dynamic environment. A further study and comparison of the detection error rate
applying this algorithm will be performed based on results. It can be concluded that the simplicity
of this algorithm allows it to be implemented in firmware and can be further fine-tuned to give
robust detection of the tag in a real environment of chipless RFID tag detection.
ACKNOWLEDGEMENTS
This work is supported by the Australian Research Council Link Project Grant
LP130101044:Discreet Reading of Printable Multi-bit Chipless RFID Tags on Polymer
Banknotes.
REFERENCES
[1] R. V. Koswatta, and N. C. Karmakar, “A novel reader architecture based on UWB chirp signal
interrogation for multiresonator-based chipless RFID tag reading,” IEEE Transactions on Microwave
Theory and Techniques, vol. 60, no. 9, pp. 2925-2933, 2012.
[2] R. V. Koswatta, and N. C. Karmakar, “Moving Average Filtering Technique for SignalProcessing in
Digital Section of UWB ChiplessRFID Reader,”Proceedings of Asia-Pacific Microwave Conference,
pp. 1304-1307, 2010.
[3] P. Kalansuriya, N. C. Karmakar, and E. Viterbo, “Signal space representation of chipless RFID
tagfrequency signatures Signal space representation of chipless RFID tagfrequency signatures,”,IEEE
Globecom 2011 proceedings, vol. 68, no. 7-8, pp. 437-445, 2013.
[4] P. Kalansuriya, N. C. Karmakar, & Viterbo, E. (2013). On the detection of chipless RFID through
signal space representation. Annales des Telecommunications/Annals of Telecommunications, 68(7-
8), 437-445. doi: 10.1007/s12243-013-0377-4
[5] R. Rezaiesarlak, & M. Manteghi, (2013). Short-time matrix pencil method for chipless RFID
detection applications. IEEE Transactions on Antennas and Propagation, 61(5), 2801-2806. doi:
10.1109/TAP.2013.2238497
[6] Rezaiesarlak, R., & Manteghi, M. (2014). A space-frequency technique for chipless RFID tag
localization. IEEE Transactions on Antennas and Propagation, 62(11), 5790-5797. doi:
10.1109/TAP.2014.2350523
[7] C. Divarathna, N.C. Karmakar, "A Maximum Likelihood Based Tag Detection Technique for MIMO
Chipless RFID Systems," IEEE International Microwave and RF Conference (IMaRC), 2014.
[8] S. Preradovic, I. Balbin, N. C. Karmakar, and G. F. Swiegers, “Multiresonator-based chipless RFID
system for low-cost item tracking,” IEEE Trans.Microw. Theory Tech., vol. 57, no. 5, pp. 1411–1419,
May 2009.

[9] A. Lazaro, A. Ramos, D. Girbau, and R.
continuous wavelet transform,” IEEE Antennas and Wireless Propag. Lett., vol. 10, pp. 520
May 2011.
[10] A. Blischak and M. Manteghi, “Pole residue techniques for chipless RFID detection,” in Proc. I
Antennas Propag. Soc. Int. Symp., Jun. 2009, pp. 1
AUTHORS
Meriam Anushani Bibile (M’16) received the BSc.Degree in Engineering Physics
from University of Colombo, Sri Lanka, in 2004 and MSc. Degree in Mobile,Personal
andSatellite Communications from University of Westminster,London, United
Kingdom, in 2006. She is currently working toward the Ph.D. degree in Electrical
Engineering at Monash University, Melbourne, Australia. From 2008
a Lecturer in Electronics and Telecommunications
University of Moratuwa, Sri Lanka. Her areas of interests include chipless RFID,
digital signal processing and digital electronics.
NemaiChandra Karmakar (S’91
information technology and electrical engineering from the University of Queensland,
St. Lucia, Australia, in 1999. He has 20 years of teaching, design, and research
experience in smart antennas, microwave ac
RFIDs in both industry and academia in Australia, Canada, Singapore and Bangladesh.
He has authored and co-authored more than 230 referred journal and conference
papers, 24 referred book chapters and three edited and
field of RFID. He has two patent applications for chipless RFIDs. Currently he is an
Associate Professor with the Department of Electrical and Computer Systems
Engineering, Monash University, Melbourne, Australia.
A. Lazaro, A. Ramos, D. Girbau, and R. Villarino, “Chipless UWB RFID tag detection using
continuous wavelet transform,” IEEE Antennas and Wireless Propag. Lett., vol. 10, pp. 520
A. Blischak and M. Manteghi, “Pole residue techniques for chipless RFID detection,” in Proc. I
Antennas Propag. Soc. Int. Symp., Jun. 2009, pp. 1–4.
(M’16) received the BSc.Degree in Engineering Physics
from University of Colombo, Sri Lanka, in 2004 and MSc. Degree in Mobile,Personal
from University of Westminster,London, United
Kingdom, in 2006. She is currently working toward the Ph.D. degree in Electrical
Engineering at Monash University, Melbourne, Australia. From 2008 – 2013 she was
a Lecturer in Electronics and Telecommunications at the Institute of Technology,
University of Moratuwa, Sri Lanka. Her areas of interests include chipless RFID,
digital signal processing and digital electronics.
(S’91-M’91-SM’99) received the Ph.D. degree in
information technology and electrical engineering from the University of Queensland,
He has 20 years of teaching, design, and research
experience in smart antennas, microwave active and passive circuits, and chipless
RFIDs in both industry and academia in Australia, Canada, Singapore and Bangladesh.
authored more than 230 referred journal and conference
papers, 24 referred book chapters and three edited and one co-authored books in the
field of RFID. He has two patent applications for chipless RFIDs. Currently he is an
Associate Professor with the Department of Electrical and Computer Systems
Engineering, Monash University, Melbourne, Australia.
Villarino, “Chipless UWB RFID tag detection using
continuous wavelet transform,” IEEE Antennas and Wireless Propag. Lett., vol. 10, pp. 520–523,
A. Blischak and M. Manteghi, “Pole residue techniques for chipless RFID detection,” in Proc. IEEE

A NOVEL ADAPTIVE - WAVELET BASED DETECTION ALGORITHM FOR CHIPLESS RFID SYSTEM

More Related Content

What's hot (20)

Similar to A NOVEL ADAPTIVE - WAVELET BASED DETECTION ALGORITHM FOR CHIPLESS RFID SYSTEM (20)

Recently uploaded (20)

A NOVEL ADAPTIVE - WAVELET BASED DETECTION ALGORITHM FOR CHIPLESS RFID SYSTEM