E3S Web of Conferences 130, 01035 (2019) [Link]
1051/e3sconf/201913001035
IC-AMME 2018
Intelligent Automatic V6 and V8 Engine Sound
Detection Based on Artificial Neural Network
Wenny Vincent1 , Astuti Winda2,* , and Mahmud Iwan Solihin3
1Automotive and Robotics Engineering Department, BINUS ASO School of Engineering, Bina
Nusantara University, Alam Sutera Boulevard no. 1, Alam Sutera-Serpong, Tangerang-Banten 15325,
Indonesia
2Computer Engineering Department, BINUS ASO School of Engineering, Bina Nusantara University,
Jl. K. H. Syahdan No. 9, Kemanggisan, Palmerah, Jakarta 11480, Indonesia
3Faculty of Engineering and Built Enviroment, UCSI University, No. 1, Jalan Menara Gading, UCSI
Heights (Taman Connaught), Cheras 56000 Kuala Lumpur, Malaysia
Abstract. The sound of V6 or V8 engines has its own cultural appeal that
cannot be replaced by the modern four-cylinder naturally aspirated or
turbocharged engines. The identification of the type of engine by the sound
is not an easy task, even for the professionals. An intelligent system that
can identify V6 to V8 engines from various cars will give an insight of the
features in the engine sounds that characterized the two different engines.
In this work, an Artificial Neural Network (ANN) approach is applied for
identifying cylinder of the engine based on the engine sound identification
is proposed. The recorded sound of the engine is then processed in order to
get some features which later be used in the proposed system. The Fast
Fourir Transform (FFT) is adopted as a feature and later used as input to
the Artificial Neural Network (ANN) based identifier. The Experimental
results confirm the effectiveness of the proposed intelligent automatic six
cylinder and eight cylinder engine based on Fast Fourier Transform (FFT)
and Artificial Neural Network (ANN), since it resulting the training and
testing accuracy of 100 % and 100 %, respectively.
Key words: Automatics petrol identification, fast fourier transform,
support vector machine.
1 Introduction
Nowadays, the development of the industrial revolution 4.0 in the autonomous vehicle
sector is experiencing rapid growth. The advance autonomous technology involve many
aspect in the vehicle, such as autonomous self driving, and full autonomous self driving [1].
One of the importance aspect in advance autonomous vehicle is engine mainatance. The
engine maintainace system plays an important rule in vehicle system, since the engine is the
main part in the automotive. Despite the common automotive trend toward smaller
displacement engines to cut down emission and fuel consumption, there are still some
market segments for high performance engine such as luxury, heavy-duty vehicles, and the
“petrohead" communities. Especially for the last one, the sound of V6 or V8 engines has its
*
Corresponding author: wastuti@[Link]
© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons
Attribution License 4.0 ([Link]
E3S Web of Conferences 130, 01035 (2019) [Link]
IC-AMME 2018
own cultural appeal that cannot be replaced by the modern four-cylinder naturally aspirated
or turbocharged engines.
The identification of the type of engine by the sound is not an easy task, even for the
professionals. An intelligent system that can identify V6 to V8 engines from various cars
will give an insight of the features in the engine sounds that characterized the two different
engines. Moreover, it can provide references for engine sound features that draw the
interest of some certain “petrol-head”communities.
Feature extraction and intelligent system are the most main stages in this work . the
sound signal is processing in the feature extraction method the feature data is the used as
input to the intelligent system. Mel-frequency cepstral coefficient (MFCC) [2], linear
predictive coding (LPC) [3], and Fast Frourier Transform (FFT) [4] are some of the
technique to extract the feature. There are many intelligent system that often applied in the
system such as support vertor machine [5], multi-layers neural network [2, 6] and fuzzy
logic [7].
In this work, an Artificial Neural Network (ANN) approach is applied for identifying
cylinder of the engine based on the engine sound identification is proposed. The recorded
sound of the engine is then processed in order to get some features which later be used in
the proposed system. The Fast Fourier Transform (FFT) is adopted as a feature and later
used as input to the Artificial Neural Network (ANN) based identifier.
2 Literature review
This section presents, the basic general process consist of two important parts, namely
sound acquisition, feature extraction and pattern matching. Sound acquisition process to
record the image by using a microphone. The next process of the recorded sound will be
extracted by feature extraction. In this process, the extracted data is the use as input to the
engine sound identification session. Finally, the sound engine recognition to make the
training process to recognize the type of engine of the automobile.
2.1 Fast Fourier Transform (FFT)
Fast Fourier Transform (FFT), is the faster Algorithm of the Discrete Fourier Transform
(DFT). Discrete Fourier transform (DFT) is a specific form of the Fourier Transform
applied to a time wave, typically a sound. Each signal is arranged from a sine / cosine
function which has a specified frequency and a relative amplitude [8].
Feature extraction process based on Fast Fourier Transform (FFT) [8], is extracted the
signal in time domain convert into frequency domain [9]. FFT is given by Nithya et al. [10],
as shown in Equation 1. The fast fourier transform of f(x) is the function of F(ω) is
explained as equation below [11];
F f x e ix dx (1)
Where i 1 and e i cos sin
2.2 Artificial neural network
Artificial Neural networks algorithm is used to simulate same or all the characteristic of the
biological neurons that form the structural constituents of the brain. Neural network can be
trained to perform a particular function by updating the value of connections (weight)
2
E3S Web of Conferences 130, 01035 (2019) [Link]
IC-AMME 2018
between elements [12]. Commonly neural networks are trained, so that particular input
leads to a specific target output. The Multi-layer perceptron, as shown in Figure 1 is
consists of three layers of neural networks; input layer, hidden layer and output layer [8].
OUTPUTs
．．． HIDDENs
INPUTs
．．．
Fig. 1. MLP neural network.
3 Proposed ANN-based engine sound identification system
This paper used ANN as engine identification based on sound. A block diagram of
conventional sound identification is shown in Figure 2. The system is trained to identify the
engine type by each person speaking out a specific utterance into the microphone. The
speech signal is digitized and some of them are carried out to create a template for the voice
pattern and stored in the memory.
Fig. 2. Block diagram of conventional engine type identification system.
Fig. 3. Overall proposed engine type identification model
As shown in Figure 3, the sound engine is recorded by the sensor. The sound signal is
then digitized in the computer. The digitaze signal is then extracted into the feature which
used as input to the identification system. FFT technique is apply as extraction method
which extract the engine signal into the numerical feature. ANN based method is applied as
identification system which idenfity the engine data feature.
3
E3S Web of Conferences 130, 01035 (2019) [Link]
IC-AMME 2018
4 Experimental Work
This section presents the performance of the proposed to automatic engine V6 and V8
engine sound identification based on FFT technique and machine learning approach. This
section consists of three important subsections, they are experimental setup, experimental
result and experimental analysis.
4.1 Experimental Design
The experimental design of engine type identification is shown in the Figure 4. The sound
of the engine is recorded using a microphone. The records sound of the engine then sent to
the PC thought 16 bits A/D with sampling frequency 44 KHz, all the signal is converted in
the PC with processor 1.5 Mhz Pentium as show in the Table 1. The sound of the engine
signal is recorded using microphone and then the recorded signal is sent to the computer
through analog to digital (A/D) card.
Analog to digital
converter
Fig. 4. Automatic V6 and V8 engine sound identification.
System identifies four different automotive engine sound from four different car with
two types of V6 and V8 engine type, respectively.
Table 1. Specification of the system.
Processor 1.5 MHz, Intel
A/D 16 bits A/D
Frequency sampling 44 KHz
4.2 Experiment Result
In order to evaluate the effectiveness of the proposed intelligent system on the cylinder
engine identification system, the proposed system is installed. In this proposes system the
specification microphones. The sound is recorded and captured. The raw recorded V8
engine signal and V6 engine as shown in Figure 5a and 5b, respectively.
Once the raw signal is recorded, then the FFT method is applied to the raw signal, as
shown in Figure 6. The FFT result of the FFT on V6 and V8 engine sound signal. From
eight sample peak amplitude from four different vehicles of each type, V6 and V8,
respectively are obtained. Four different samples of each vehicle of different type of engine
are used to train the system. Furthermore, the other four different samples of each of the
vehicles samples are applied to test the system. In this experiment The ANN-based binary-
class classification is applied to perform the classification process of the training phases.
The training classification rate for these two features based resulting 100 % correct and the
testing classification rate lies at 100 %.
4
E3S Web of Conferences 130, 01035 (2019) [Link]
IC-AMME 2018
(a) Raw signal of V6 engine
(b) Raw signal of V8 engine
Fig. 5. Raw V6 and V8 signal.
(a) FFT result of the V6 engine signal
5
E3S Web of Conferences 130, 01035 (2019) [Link]
IC-AMME 2018
(b) FFT result of the V8 engine signal
Fig. 6. FFT result of the V6 and V8 engines.
4.3 Experiment analysis
The distance between engine and microphone, noise environments and the stability of the
sounds of the engine are affected with the recording quality. These factors impact the FFT
algorithm. Table 2 discusses the accuracy between different types of V6 and V8 engine
based on the sound of the engine. FFT algorithm feature extraction has applied as the
feature extraction method. The result shows that the system has good accuracy, since it has
100 % accuracy for both training and testing phases, respectively.
Table 2. Experimental results of training and testing phase.
Training Testing
Type Classification
(Classification rate %) (Classification rate %)
V6 engine type 100 100
V8 engine type 100 100
5. Conclusion
The sound of V6 or V8 engines has its own cultural appeal that cannot be replaced by the
modern four-cylinder naturally aspirated or turbocharged engines. The identification of the
type of engine by the sound is not an easy task, even for the professionals. An intelligent
system that can identify V6 to V8 engines from various cars will give an insight of the
features in the engine sounds that characterized the two different engines.
In this work, an Artificial Neural Network (ANN) approach is applied for identifying
cylinder of the engine based on the engine sound identification is proposed. The recorded
sound of the engine is then processed in order to get some features which later be used in
the proposed system. The Fast Fourier Transform (FFT) is adopted as a feature and later
used as input to the Artificial Neural Network (ANN) based identifier. ANN-based engine
sounds identification has been proposed and discuss in this paper. The significant FFT
frequencies used as features. The result shows that FFT frequencies have successfully
performed as good features. The accuracy of the ANN based engine sound identification
using FFT frequencies is 100 %.
6
E3S Web of Conferences 130, 01035 (2019) [Link]
IC-AMME 2018
References
1. Asean Development Bank. What does the Fourth Industrial Revolution mean for
regional economic integration ? Geneva: World Economic Forum (2017).
[Link]
[Link]
2. N. Dave. Int. J. Adv. Res. Eng. Technol., 1,VI :1–5(2013).
[Link]
3. Thiang, S. Wijoyo. Speech recognition using linear predictive coding and artificial
neural network for controlling movement of mobile robot. International Conference on
Information and Electronics Engineering (Bangkok, Thailand, 2011). IPCSIT, 6:179–
183(2011). [Link]
4. H. Ali, M.J.E. Salami, Wahyudi. Iris recognition system by using support vector
machines. Proceedings of the International Conference on Computer and
Communication Engineering 2008, ICCCE08: Global Links for Human Development
(Kuala Lumpur, Malaysia, 2008). [Link]
5. T. Seehapoch, S. Wongthanavasu. Speech emotion recognition using support vector
machines. The 5th International Conference on Knowledge and Smart Technology
(KST), (Chonburi, Thailand, 2013). p. 86–91.
[Link]
6. P. Potocnik. Neural networks: MATLAB examples. Slovenia: University of Ljubljana
(2012).
[Link]
Networks_course_practical_examples_2012_Primoz_Potocnik
7. Wahyudi, W. Astuti, S. Mohamed. Journal of Computer Science, 3,5:274–280(2007).
[Link]
8. N.T. Deshpande, S. Ravishankar. Adv. Comput. Sci. Technol., 10,5:1173–1189(2017).
[Link]
9. D. Gunawan, F.H. Juwono. Pengolahan Sinyal Digital dengan Pemograman Matlab.
[Digital Signal Processing with Matlab Programming] Yogyakarta: Graha Ilmu (2012).
[in Bahasa Indonesia]. [Link]
pemrograman-matlab/
10. A. Nithya, C. Lakshmi, R. Krithiga (2017). IJPAM, 113,12: 57–66(2017).
[Link]
CM_and_FFT_Feature_Set_Fusion
11. J. Sueuer. A very short introduction to sound analysis for those who like elephant
trumpet calls or other wildlife sound. Paris: Mus´eum national d’Histoire naturelle
(2018). [Link]
[Link]/web/packages/seewave/vignettes/seewave_analysis.pdf
12. J.C. Russ. The Image Processing Handbook. Florida: CRC Press (2011).
[Link]
