Quartz crystal microbalance based electronic nose system implemented on Field Programmable Gate Array

TELKOMNIKA, Vol.17, No.1, February 2019, pp.370~376
ISSN: 1693-6930, accredited First Grade by Kemenristekdikti, Decree No: 21/E/KPT/2018
DOI: 10.12928/TELKOMNIKA.v17i1.10133  370
Received June 2, 2018; Revised October 10, 2018; Accepted November 12, 2018
Quartz crystal microbalance based
electronic nose system implemented on
Field Programmable Gate Array
Misbah
1
, Muhammad Rivai*
2
, Fredy Kurniawan
3
1,2
Department of Electrical Engineering, Institut Teknologi Sepuluh Nopember
Arief Rahman Hakim St., Keputih, Sukolilo, Surabaya, Indonesia
3
Department of Chemistry, Institut Teknologi Sepuluh Nopember
3
ITS Halal Center, Institut Teknologi Sepuluh Nopember
Arief Rahman Hakim St., Keputih, Sukolilo, Surabaya, Indonesia
1
Department of Electrical Engineering, Muhammadiyah University of Gresik
Sumatera 101 GKB St., Kebomas Gresik, Indonesia
*Corresponding author, e-mail: muhammad_rivai@ee.its.ac.id
Abstract
Nowadays, an electronic nose becomes an important tool for detecting gas. The electronic nose
consists of gas sensor array combined with neural networks to recognize patterns of the sensor array.
Currently, the implementation of the neural network on the electronic nose systems still use personal
computer so that less flexible or not portable. This paper discusses the electronic nose system
implemented in a Field Programmable Gate Array (FPGA). The sensor array consists of eight Quartz
Crystal Microbalance (QCM) coated with chemical materials. The eight channel-frequency counter is used
to measure the frequency change of the sensor due to the presence of gas adsorbed to the surfaces. The
bipolar sigmoid activation function used in the neuron model is approximated by a second order equation.
The experimental result showed that the electronic nose system could recognize all the types of gas with
92% success rate.
Keywords: electronic nose system, FPGA, neural network, QCM
Copyright © 2019 Universitas Ahmad Dahlan. All rights reserved.
1. Introduction
The electronic nose system has been widely applied in broad variety of fields, such as
for monitoring coffee quality [1], plant disease diagnosis [2], fruit classification [3] and air
monitoring [4]. This system consists of a sensor array combined with pattern recognition
algorithm to identify the gas. Currently, electronic nose systems still use computers [5], which
are less flexible and are not portable.
In recent years, there have been studied several gas sensors with different working
principles, such as electro-chemical sensors [6], metal oxide semiconductor sensors [7, 8], ratio
metric fluorescence sensors [9], fiber optic sensors [10, 11], and Quartz Crystal Microbalance
(QCM) [12, 13]. Compared with other gas sensors, QCM sensors are simple, cheap, small, and
high sensitive to the weight of molecules.
Field Programmable Gate Array (FPGA) has recently expanded the structure with the
presence of digital signal processing [14, 15]. FPGA technology makes it easy to reconfigurable,
reprogrammable with low cost advantages, and easy to integrate [16]. In many applications,
FPGA can be used in implementing pattern recognition algorithms, such as SOM-Hebb [17],
Learning Vector Quantization (LVQ) [18, 19], and neural networks [20, 21]. In neural network
architecture, there are several activation functions, such as sigmoid equation. When
implementing the sigmoid function in the FPGA, it often has constraints in hardware
implementation. Some previous studies have approached the sigmoid function, such as with
Lookup Table (LUT) combinations using linear interpolation [22], exponential function [23], the
4
th
order Taylor series [24], and second order function [25].
Several studies have implemented electronic nose on FPGA, including the
implementation of metal oxide semiconductor gas sensors equipped with Multilayer perceptron
(MLP) with single multiple inputs multiple outputs (SMIMO) and multiple multiple inputs single

TELKOMNIKA ISSN: 1693-6930 
Quartz crystal microbalance based electronic nose system implemented... (Misbah)
371
output (MMISO) [26]. Another study uses gas sensor array with principal component analysis
(PCA) as a pattern recognition method [27]. In this study, we have implemented electronic nose
on FPGA using polymer coated QCM sensor array. Some of the gases used as sample material
are kerosene, gasoline, ammonia, and ethanol. The second order function model is used as an
activation function in the neural network pattern recognition.
2. Research Method
Quartz resonator sensor has high sensitivity to the gas molecules adsorbed to the
electrodes shown in Figure 1. This will change the oscillation frequency of the quartz resonator,
defined as (1):
m
vA
f
f
SiO










2
2
2

(1)
where Δf is the observed frequency shift (Hz), Δm is the mass change per unit area (g/cm
2
), ρ is
the crystal density, v is the velocity of acoustic wave propagation in quartz crystal, A is the
surface area (cm
2
).
Figure 1. The basic principle of QCM sensor
The design of the electronic nose consists of an array of QCM sensors, oscillator
circuits, 1-second time base, frequency counters, latches, neural network, and LCD driver
shown in Figure 2.The Pierce oscillator accompanied by a QCM sensor is used as a gas sensor
unit. There are eight QCMs coated with various chemicals including OV-17, PEG-20M, OV-225,
OV-25, PEG-200, OV-101, PEG-1540, and PEG-6000. The time base of one second is obtained
from the delay process for the 50 Mhz oscillator circuit attached on the FPGA module.
By using a time delay of one second, it allows to measure the frequency change with a
resolution of 1 Hz. The frequency counter block can be constructed from several flip-flop
circuits, such as D, RS, or JK Flip-flops. In VHDL code, frequency counter is designed with
24 bits using behaviour architecture.
The neural network architecture used in this study consists of three layers shown in
Figure 3. This network consists of 8 nodes in the input layer, 10 neurons in the hidden layer,
and 2 neurons in the output layer. Generally, the supervised neural network algorithms are
using MLP with back propagation learning method. The hidden neuron is expressed by:


n
i
ijijj vxvnetz
1
0_
𝑧𝑗 = 𝑓1(𝑧_𝑛𝑒𝑡 𝑗) =
1−𝑒
−𝑧_𝑛𝑒𝑡 𝑗
1+𝑒
−𝑧_𝑛𝑒𝑡 𝑗
(2)
the neuron output is expressed by:


p
j
jkjkk wzwnety
1
0 ._
𝑦 𝑘 = 𝑓2(𝑦_𝑛𝑒𝑡 𝑘) =
1−𝑒−𝑦_𝑛𝑒𝑡 𝑘
1+𝑒−𝑦_𝑛𝑒𝑡 𝑘
(3)

 ISSN: 1693-6930
TELKOMNIKA Vol. 17, No. 1, February 2019: 370-376
372
Figure 2. Block diagram of the electronic nose system used in the experiment
Figure 3. The feedforward neural network architecture
Xilinx ISE software is used in synthesis and compilation of VHDL code. Selection of
data types affects the compilation performance. The compilation process will have many
problems when the (2) and (3) are implemented in VHDL code. The second order equation is an
easy approach to replace non-linear functions such as the sigmoid function. The transition from
the lower and upper saturation regions of the sigmoid function can be
replaced by [25]:
𝐻(𝑧) = {
𝑧(𝛽 − 𝜃𝑧)𝑓𝑜𝑟 0 < 𝑧 < 𝐿
𝑧(𝛽 + 𝜃𝑧)𝑓𝑜𝑟 − 𝐿 < 𝑧 < 0
} (4)
where, β and θ are used to adjust the slope and gain of non-linear functions. The equation can
be written as:
𝐺(𝑧) = {
1 𝑓𝑜𝑟𝑧 > 𝐿
𝐻(𝑧)𝑓𝑜𝑟 − 𝐿 < 𝑧 < 𝐿
−1 𝑓𝑜𝑟𝑧 < −𝐿
} (5)
to approximate the sigmoid equation, we get θ = 1/L
2
and β = 2/L, with L = 4 and then a slightly
modify the equation to be:

373
𝐺(𝑧) = {
1 𝑓𝑜𝑟 𝑧 > 4
𝐻(𝑧)𝑓𝑜𝑟 − 4 < 𝑧 < 4
−1 𝑓𝑜𝑟𝑧 < −4
} where, 𝐻(𝑧) = {
0.52 𝑧 − 0.07 𝑧2
𝑓𝑜𝑟 0 < 𝑧 < 4
0.52 𝑧 + 0.07 𝑧2
𝑓𝑜𝑟 − 4 < 𝑧 < 0
} (6)
An integer type is suggested in this design to approach floating-point operation.
Therefore, all numbers need to be changed in the form of integer values with multiplying by
1024, which the steps can be written as:
a. Normalization of the inputs x1, x2, ... xn are represented by:𝑥 𝑛𝑜𝑟𝑚 =
𝑥×1024
𝑥 𝑚𝑎𝑥
b. The weight and bias values v11, v12...vij; w11, w12... wjk are multipled by 1024.
c. The hidden neuron: 𝑧 𝑛𝑒𝑡 =
∑ 𝑖𝑛𝑝𝑢𝑡(𝑥 𝑛𝑜𝑟𝑚)×𝑤𝑒𝑖𝑔ℎ𝑡𝑠(𝑣 𝑖𝑗)
1024
+ 𝑏𝑖𝑎𝑠
𝑧 = 𝐺(𝑧 𝑛𝑒𝑡)
with the sigmoid equation (G) is represented by:
𝐺(𝑔 𝑛𝑒𝑡) =
{
1024 ; 𝑓𝑜𝑟𝑔 𝑛𝑒𝑡 > 4000
(532×𝑔 𝑛𝑒𝑡−
72×𝑔 𝑛𝑒𝑡
2
1024
)
1024
; 𝑓𝑜𝑟 0 ≤ 𝑔 𝑛𝑒𝑡 < 4000
(532×𝑔 𝑛𝑒𝑡+
72×𝑔 𝑛𝑒𝑡
2
1024
)
1024
; 𝑓𝑜𝑟 − 4000 ≤ 𝑔 𝑛𝑒𝑡 < 0
−1024 ; 𝑓𝑜𝑟𝑔 𝑛𝑒𝑡 < −4000 }
(7)
d. The output neuron: 𝑦 𝑛𝑒𝑡 =
∑ 𝑧×𝑤𝑒𝑖𝑔ℎ𝑡𝑠 (𝑤 𝑗𝑘)
1024
+ 𝑏𝑖𝑎𝑠
𝑦 = 𝐺(𝑦 𝑛𝑒𝑡)
the VHDL code of the neurons can be expressed as follows:
the VHDL code for the LCD driver follows the LCD vendor's instructions. Hitachi LCD is used on
this system with the two 4-bit operations of data transfer. In order to connect to the Input/Output
ports of the FPGA, the result of the each block must be converted into a standard logic vector
type.
3. Results and Analysis
The electronic nose system has been implemented on Spartan
3E XC3S1600E-5FG320 starter kit board.The FPGA module consists of two RS232 ports, a
VGA port, 2x16 LCD, 10/100 Ethernet PHY, a 40 pin Hirose FX2 connector, 29,504 flip flops,
29.504 LUTs, 32M x 16 DDR SDRAM and thirty-six of multiplier 18x18.Prior to applying in the
real system, the frequency counter is simulated on Xilinx ISE Software. The measurement of the
frequency counter hardware is performed for 8-channels. The experimental result shows that
this counter unit has an inter-channel cross talk of 0.0002%.The more channels there will be
more interference between channels. In previous study, it was only used 2-channels with
precision and stability of 1 Hz with TCXO oscillator [28].
Figure 4 (a) shows a comparison between the actual bipolar sigmoid and the second
order functions; while Figure 4 (b) shows the bipolar sigmoidmultiplied by 1000 and the second
order multiplied by 1024. The average error for the second order and the modified second order
are 1.7% and 2.4%, respectively, as shown in Figure 4 (c). The results of the implementation of
the second order function indicate that the number of Slices is 143, the number of 4 inputs of
LUTs is 274, and the number of IOBs is 10. The summary of device utilization is still greater at
27%, and 31% for slices and LUTs, respectively, due to the implementation of LCD drivers.
However, the use of 84% of IOBs is less than previous study using fixed-point arithmetic
simulated in ModelSim [29].
if ( znet< 0 and znet> -4000) then z<= (532*znet + (72*znet*znet)/1024)/1024;
elsif (znet< -4000) then z<= -1024;
elsif (znet>= 0 and znet< 4000) then z<= (532*znet - (72*znet*znet)/1024)/1024;
else z<= 1024;
end if;

 ISSN: 1693-6930
374
(a)
(b)
(c)
Figure 4. The comparison of sigmoid function with (a) second order,
(b) modified second order, and (c) error plots of the both approaches
For data analysis, this system requires a computer to display the response of each
sensor in real time shown in Figure 5. Figure 6 shows the patterns of the sensor array for any
gases with different concentrations. The results of this experiment indicate the similarity of
sensor pattern under varied gas concentration. Each gas provides a typical pattern of sensor
response. These response patterns are then used as inputs on the neural network to recognize
the type of gas.

375
Figure 5. The time response of QCM sensors exposed to ammonia vapor
The training phase of the neural network is performed on a computer using
backpropagation algorithm. For the iterations of 100,000 epochs, the network has a Mean
Square Error (MSE) of 0.00005. The weights of the training phase are entered into the FPGA.
The overall use of components in Slice of Flip Flop is 71% of 29,040; 4 inputs LUT is 68% of
29,040; Multiplier 18x18 is 75% of 36; and buffermultiplex is 41% of 24. In the testing phase, the
neural network is introduced with forty data sets for all types of gases. The experimental result
shows that the neural network has an average identification rate of 92% as shown in Table 1.
Figure 6. The patterns of sensor responses for several gases with varied concentrations
Table 1. The Identification Rate of the Neural Network
No Gas Samples Success Rate (%)
1 Gasoline 100
2 Ammonia 100
3 Kerosene 80
4 Ethanol 90
Average 92
4. Conclusion
The electronic nose system based on a QCM sensor array and neural network pattern
recognition has been constructed and implemented on the XC3S1600e FPGA board. The
bipolar sigmoid activation function used in the neuron model is approximated by a second
orderfunctionwith an error rate of 2.4%. The overall design of the system requires the use of

 ISSN: 1693-6930
376
slices and LUTs of 71%, and 68%, respectively. The results of the experiment indicate the
similarity of sensor pattern to varied gas concentration. Each gas provides a typical pattern of
sensor response. The neural network could recognize all the gases tested in the experiments
with an identification rate of 92%.The use of the counter circuits in this experiment requires a
sampling time of 1 second. To increase the respond time, a reciprocal frequency counter will be
developed in this electronic nose system. The graphical LCD will be implemented to display the
data in detail.
Acknowledgment
This research was carried out with financial aid support from the Ministry of Research,
Technology and Higher Education of the Republic of Indonesia (Kemenristekdikti RI).
References
[1] Radi, Rivai M, Purnomo MH. Study on Electronic-Nose-Based Quality Monitoring System for Coffee
Under Roasting. Journal of Circuits, Systems, and Computers. 2016; 25(10): 1650116. Available
from:http://www.worldscientific.com/doi/abs/10.1142/S0218126616501164.
[2] Cellini A, Blasioli S, Biondi E, Bertaccini A, Braschi I, Spinelli F. Potential Applications and Limitations
of Electronic Nose Devices for Plant Disease Diagnosis. Sensors. 2017; 17(11): 2596. Available
from: http://www.mdpi.com/1424-8220/17/11/2596.
[3] Adak M, Yumusak N. Classification of E-Nose Aroma Data of Four Fruit
Types by ABC-Based Neural Network. Sensors. 2016; 16(3): 304.
Available from: http://www.mdpi.com/1424-8220/16/3/304.
[4] He J, Xu L, Wang P, Wang Q. A high precise E-nose for daily indoor
air quality monitoring in living environment. Integration. 2017; 58: 286–294. Available from:
https://www.sciencedirect.com/science/article/pii/S0167926016301985.
[5] Xu M, Wang J, Gu S. Rapid identification of tea quality by E-nose and computer vision combining
with a synergetic data fusion strategy. J Food Eng. 2019; 241: 10–17. Available
from:https://www.sciencedirect.com/science/article/pii/S0260877418303091.
[6] Fitriyana F, Kurniawan F. Polyaniline-Invertase-Gold Nanoparticles Modified Gold Electrode for
Sucrose Detection. Indones J Chem. 2015; 15(3): 226–233. Available from:
http://pdm-mipa.ugm.ac.id/ojs/index.php/ijc/article/view/ 1000.
[7] Monroy JG, Gonzalez Jimenez J. Gas classification in motion: An experimental
analysis. Sensors Actuators, B Chem. 2017; 240: 1205–1215. Available from: http://
dx.doi.org/10.1016/j.snb.2016.09.013.
[8] Sinha SK. Growth and ammonia sensing properties of Zn1−xSnxO nanofibers.
Sensors Actuators B Chem. 2015; 219: 192–198. Available from: https://www.sciencedirect.com/
science/article/pii/S092540051500619X?via%3Dihub.
[9] Jazan E, Mirzaei H. Direct analysis of human breath ammonia using corona discharge
ion mobility spectrometry. J Pharm Biomed Anal. 2014; 88: 315–320. Available
from:https://www.sciencedirect.com/science/article/abs/pii/ S073170851300410X.
[10] AL Khalaf, FS Mohamad, et al. Room temperature ammonia sensor using side-polished optical fiber
coated with graphene/polyaniline Nanocomposite. Optical Materials Express. 2017; 7(6):
1858-1870.
[11] Ibrahim SA, Rahman NA, Abu Bakar MH, Girei SH, Yaacob MH, Ahmad H, et al.
Room temperature ammonia sensing using tapered multimode fiber coated
with polyaniline nanofibers. Opt Express. 2015; 23(3): 2837. Available from:
https:// www.osapublishing.org/abstract.cfm?URI=oe-23-3-2837.
[12] Bearzotti A, Macagnano A, Papa P, Venditti I, Zampetti E. A study of a QCM sensor based on
pentacene for the detection of BTX vapors in air. Sensors Actuators, B Chem. 2017; 240:
1160–1164. Available from: http://dx.doi.org/10.1016/j.snb.2016.09.097
[13] Wang L, Wang Z, Xiang Q, Chen Y, Duan Z, Xu J. High performance formaldehyde detection based
on a novel copper (II) complex functionalized QCM gas sensor. Sensors Actuators, B Chem. 2017;
248: 820–8. Available from: http://dx.doi.org/10.1016/j.snb.2016.12.015.
[14] Bernard Schwarz M, Zwick W, Wenzel L, Klier J, Gröschl M. Field programmable gate array-assigned
complex-valued computation and its limits. Rev Sci Instrum. 2014; 85(9): 93104. Available from:
http://aip.scitation.org/doi/10.1063/1.4894211
[15] Schwettmann A, Sedlacek J, Shaffer JP. Field-programmable gate array based locking circuit for
external cavity diode laser frequency stabilization. Rev Sci Instrum. 2011; 82(10): 103103. Available
from: http://aip.scitation.org/doi/10.1063/1.3646477.
[16] Jinghong L, Kai L, Peng Y. Design and implementation of flame combustion state detection system
based on FPGA. In: The 27th Chinese Control and Decision Conference (2015 CCDC). IEEE; 2015:
1392–1396. Available from: http://ieeexplore.ieee.org/document/7162136/

Quartz crystal microbalance based electronic nose system implemented on Field Programmable Gate Array

More Related Content

What's hot (20)

Similar to Quartz crystal microbalance based electronic nose system implemented on Field Programmable Gate Array (20)

More from TELKOMNIKA JOURNAL (20)

Recently uploaded (20)

Quartz crystal microbalance based electronic nose system implemented on Field Programmable Gate Array