Real Time Monitoring System for Automotive Tire Set using an
Acoustic Signal
C. J. Chen
1
, L. S. Chao
1
, T. H. Li
2
, C. W. Hsu
2
, S. K. Chiu
2
, C. H. Su
3
, L. H. Huang
3
, S. Y. G. Tai
3
,
Y. C. Chuang
4
, C. S. Gong
4
, C. W. Su
5
and F. Y. Su
5
1
Department of Engineering Science, National Cheng-Kung University, Tainan City, Taiwan
2
Chiu Chau Enterprise Co., LTD, Taoyuan City, Taiwan
3
Signal Technology Instrument Inc., Taoyuan City, Taiwan
4
Department of Electrical Engineering, Chang Gung University, Taoyuan City, Taiwan
5
Chain Young Co., Ltd, NewTaipei City, Taiwan
Key
words: Automotive Tire Set, Oder Tracking Technique, Acoustic Signal.
Abstract: In this research, we present a real time monitoring system for Automotive Tire Set (ATS) using an acoustic
signal. Order tracking techniques applied in signal processing module by using the recursive Kalman
adaptive filtering algorithm which can be used to exact the order amplitudes of global acoustic signals. The
detective system combines a signal processing module and orders amplitude calculated of feature extraction.
On the basis of acoustic signal extraction to distinguish between normal and abnormal types can be used to
acquire high resolution amplitude of different orders. The diagnostic process is based on recording the drive
shaft of tire speed. This system is implemented on the platform of National Instruments (NI) compact-RIO
and c-series modules. Instantaneous shaft speed and acoustic signal are respectively detected by means of
fiber optical and array microphone. Experimental results were carried out for a practical Mitsubishi Freeca
Car to appear obviously the effectiveness of the proposed system in tracking the ATS orders with high
recognition rate. An intelligent prediction integration system with internet (IPII) for ATS was proposed.
1 INTRODUCTION
In recent years, issues like green energy and Internet
of Vehicles (IoV) (He et al., 2014) developed with a
large amount sensors to detect the predictive faults.
The firms may achieve advantages of fast
production, small-volume production of a wide
range of different items, and great cost reduction,
etc., which are the promotion cores of Taiwan
government’s Productivity 4.0 plan The IoV concept
is also included in it, they called Internet of
Everything (IoE). The vehicles are now becoming
more and more complex with increased reliability on
electronics and on-board computers. Hence, fault
diagnosis on engines of vehicle has become
increasingly challenging with a great number of
parts and systems. Therefore, the job of fault
diagnosis of vehicle has become more difficult,
particularly for nonroutine faults. Automotive
manufacturers develop a new electronic diagnostic
systems which can help lead quickly to the root of
vehicle faults. Because of the limited resources of a
vehicle (less information storage and slow
processor) for electronic diagnostic apparatus, it is
difficult to do much more than limit type
diagnostics. Advanced signal processing techniques
are employed and expanded as signal
transformations or machine learning techniques.
Generally speaking, procedure of operating on
internal combustion (I.C.) engine belongs to rotating
machinery in mechanical system. The classifications
of fault in rotating mechanical system are resonance,
bearing fault, power unbalance, gear fault, and so
forth. Recently, fault diagnosis of rotating
machinery in automobile is based on shaft speed
signal and vibration signal to monitor the conditions
of system. The vibration energy or pressure energy
are often exploited in fault diagnostic system which
can be monitored the conditions of vehicle engine
when damage is increased. On the other hand, the
acoustic energy produced by rotating machine
usually reflects the working condition. An
168
Chen, C., Chao, L., Li, T., Hsu, C., Chiu, S., Su, C., Huang, L., Tai, S., Chuang, Y., Gong, C., Su, C. and Su, F.
Real Time Monitoring System for Automotive Tire Set using an Acoustic Signal.
In Proceedings of the International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2016), pages 168-173
ISBN: 978-989-758-185-4
Copyright
c
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
experienced technician can make the troubleshooting
by means of listening to the sound of machine. A
diagnostic system using acoustic energy will directly
diagnose in rotating machine. In order to obtain the
useful information by using an acoustic energy for
feature extraction is based on signal processing
techniques when the heavy noise is increased around
the rotating machinery.
The features of fault diagnosis in an I.C. gasoline
engine can be monitored by measuring the vibration
signal, acoustic signal or pressure signal at specific
locations, e.g. engine mounts. Among these signals
mainly comprised a basic frequency with related to
harmonic frequencies, most of which correlates with
revolution of crankshaft of engine. The acoustic
energy or vibration energy is exceptionally increased
when the translation systems are harmed. In general,
a traditional diagnosis method is used to observe the
amplitude difference of energy in time-frequency
domain for fault diagnosis. For instant, the fast
Fourier transform (FFT) methods are used to
observe the amplitude difference in time-frequency
domain at the fixed speed. However, the information
thus obtained is only partial because some features
of fault do not respond significantly at the fixed
operation speed. Hence the smearing problem
generally arises in practical implementation
particularly at various speeds.
A well-known approach is also utilized, the order
tracking technique that exploits acoustic energy or
vibration signals, supplemented with information of
crankshaft speed, serves as a useful tool for
diagnosis of the vehicle engine. The comparison
between frequency analysis and order analysis is
summarized in Table 1. In general, conventional
methods of order tracking are mainly based on a
Fourier analysis. A high resolution order tracking
technique is proposed in order to overcome the
smearing problems arising. Recently, fault diagnosis
of order tracking technique has become one of the
significant approaches in rotating machinery. Using
the order tracking technique can provide the feature
of order spectra from vibration signal and shaft
speed for an I.C. engine. Moreover, an order
spectrum gives the amplitude of signal as a function
of harmonic and crankshaft speed (R.P.M.). Order
tracking is mainly used to analyze and track the
energy of order signal from dynamic signal.
However, generally tracking methods are ineffective
for applications including the multiple independent
shaft speeds. For instance, the shaft speed of an I.C.
engine and the speed of cooling fans are
independent. If one calculates the orders based on
either speeds, the signals related to other speeds
would appear as uncorrelated noise and reduced the
tracking accuracy of the results. In order to avoid
aforementioned problems encountered, the
representative model-based methods have been
proposed, such as an adaptive Kalman filtering
method. In 1996, an adaptive filter theory is
published by Haykin (Haykin, 2002), discussed
some conclusions for adaptive filtering
methodologies and order tracking techniques in fault
diagnosis of rotating machinery.
In this work, a high resolution order tracking
fault diagnosis technique (Wu, et al., 2009) is
proposed for tracking the signals of acoustical
energy of an ATS set. This technique exploits
adaptive filtering based on the Kalman filter
algorithm, the proposed methods also requires the
information of shaft revolution of ATS and
measured by a fiber optical sensor. In this technique,
order amplitudes are calculated off-line by using a
least-squares approach. The adaptive algorithm is
essentially sample-based and the order amplitudes
can be calculated in a real-time fashion for fault
diagnosis of ATS. The technique is implemented on
a NI cRIO 9075 platform for evaluating the
performance in practically diagnostic systems. On
the other hand, the information of vibration signal is
the most widely used in the diagnostic analysis of
fault. In the case of fault diagnosis application
systems by using the vibration reference signal may
not be available. As a result of effect of uncertain
conditions around the vehicle ATS can likely be
generated as the additional vibration delivered from
the ground. Thus, a sound acoustical signal provides
accurate information to the fault diagnosis system.
This research scope is mainly a solution for ATS,
which integrates the sensing and predicting
technology of rotary equipment, order analysis
theory, NI cRIO embedded hardware, Real-Time
Module, FPGA, and PC-Based Guide User Interface,
etc., and the major purpose is to provide the real-
time statuses of rotary equipment. This paper issued
an Intelligent Prediction Integration System with
Internet, IPII. ATS set play a significant role in
vehicle, therefore a sudden broken is unfavourable;
and the intelligent sensing system application can
not only maintain the operating conditions at the
best status to hold life of people. Figure 1 is the
simulation fixture of an intelligent sensing module
system.
For the proposed method, a sound acoustical
signal is exploited to evaluate the proposed
algorithms. The results indicate that the proposed
method is well suited for the tracking of closely
spaced orders or crossing orders without significant
Real Time Monitoring System for Automotive Tire Set using an Acoustic Signal
169
smearing problems. Experiments are carried out to
evaluate the proposed system with practically
running tests under fixed revolution conditions The
filtering algorithms and the order tracking
techniques will be described.
Table 1: Relationship of frequency and order analysis.
Frequency analysis Order analysis
Time [sec] Revolutions [rev]
Frequency [Hz]
[per sec]
Harmonics [Order]
[per revolution]
Figure 1: Simulation fixture of ATS.
2 THEOREY OF ORDER
TRACKING
The speed, acoustics, and vibration signals of rotary
machinery must be calculated, then one transfers the
discrete time signals to discrete angle signals.
Discrete time signals are acquired according to equal
time intervals while discrete angle signals are
sampled by equal shaft angles, and this is the so-
called Resampling theorem. Angle-Samples can be
acquired by hardware devices or software post-
processing.
The hardware solution can achieve the equal
angle interval sampling purpose by using Encoder or
Tachometer’s impulse signals to trigger Analog-to-
Digital Conversion. And this sampling method
requires extra hardware devices and complex-
numbered analog filters, which is not easy to
implement. Hence this paper adopted the software
post-processing method for the experiment. The
method first collects vibration signals and
tachometer’s impulse signals at the same time to
interpolate or Curve-Fitting tachometer’s impulse
signals to receive relative angular displacements at
each time point, then to sample vibration and
acoustic signals based on the equal angular
displacement principle.
The resampled signal are called Order Signal,
and the spectrogram acquired from having FFT or
STFT on Order Signal is called Order Spectrogram
which enables one to easily tell Fundamental
Frequency and Harmonies from the analyzed signals,
and helps to analyze acoustic signal features.
2.1 Kalman Filtering
The signals generated by rotary machinery are
Frequency-Modulated Signal, which usually
contains various frequency elements (main rotary
frequency or its harmony wave). Therefore, signals
can be presented by Superposition with sinusoids of
different frequencies, and we used adaptive Kalman
filter to implement big data computing of rotary
equipment, and the Data Equation is the part of
sensed speed, acoustics data or estimated energy, the
flat condition of estimating order is called Structural
Equation. An Kalman filter algorithm presented
solutions by using State-Space, and is also able to be
resolved by recursive method.
The acoustic signal
()
x
t
containing k orders
generated by the rotating shaft from an engine
platform can be expressed as
112 2
() cos[ () ] cos[2 () ] cos[ () ]
kk
xt A t A t A k t


(1)
where A
k
and
k
express the amplitude and phase,
respectively, of the kth order, and the variable
()t
denotes the angular displacement of the shaft
computed by the following integral
00
() ( ) 2 ( )
tt
tdfd



(2)
where
()
represents the instantaneous angular
frequency (rad/sec) of the ATS shaft which should
be calculated by numerically integrating the
instantaneous angular frequency measured by the
fiber optical sensor from ATS shaft. The block
diagram representation of the adaptive Kalman
filtering is shown as Figure 2, which the procedure
is summarized as follows
Figure 2: Block diagram of Adaptive Kalman filtering.
)(ny
)1(
nx
)(
1
nv
)(nx
)(
2
nv
I
1
z
)(nC
),1( nn
Φ
VEHITS 2016 - International Conference on Vehicle Technology and Intelligent Transport Systems
170
The procedure is summarized as follows
)(),1()(),1(),1(
1
nnnnnnnn
H
QFKFK
(3)
The equation (3) can be represented a Riccati
equation solver for computing the updated value
(1,)nnK
. To initialize the Kalman filtering process,
the initial conditions are generally taken to be:
1)24(0
)y|1(
ˆ
k
0x
(
42k
is the number of
parameters
)(nA
i
),
I
K
)0,1(
(
I
is a
(
42k
)
(
42k
) identity matrix. Thus in
applying this order tracking method, different shaft
speed of axles must be available. However, the prior
information of the number of order and resonance
frequency is also required. Technically, this can be
obtained by a preliminary scan by using
conventional order tracking methods. If this can be
done, the proposed technique should provide results
with improved accuracy for prediction system on a
ATS. The IPII monitoring structure of ATS is shown
as figure 3.
Figure 3: Structure of an IPII.
3 RESULT AND VERIFICATION
The research is mainly based on applying multi-
sensor big data algorithm, and uses acoustic, speed,
and temperature, etc. to sense signals to extract and
analyze then determine, figure 4 shows the
intelligent ATS experimental structure which
includes sensors from ATS system, fiber optical
tachometer to microphone and the system module is
an embedded 667MHz CPU, with an adaptive
Kalman filter for kernel algorithms.
The IPII system was set to sample frequency at 5
KHz; the extraction time was 12 seconds; FFT
Record Length was 2,048 points; Hanning window
function. Figure 5 is the user interface which can
Figure 4: Intelligence ATS structure.
Figure 5: IPII GUI.
operate the ATS equipment on CPU end, and is able
to link to tablets and continue serving. Figure 6
shows the fiber optical from the ATS condition, the
on-line microphone measures the feature shown as
figure 7-8. Figure 9 is shown as the time signal of
practical microphone.
All the sensors with the network device, as well
as a microphone to perform big data real-time
computing, there was an axle failure, thus figure 10
shows the big data computing result of the acoustic
energy; the solid line is about normal energy
distribution condition of the ATS condition.
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
-5
-4
-3
-2
-1
0
1
2
3
4
Figure 6: ATS signal from fiber optical.
Real Time Monitoring System for Automotive Tire Set using an Acoustic Signal
171
This paper applied an adaptive algorithm, and
with the network device, one can learn real-time
condition of the current rotary equipment, then to
monitor, predict, and adjust the equipment status in
time, which also meets the predicting and sensing
technology of Taiwan’s Productivity 4.0 plan.
0 1 2 3 4 5 6
x 10
4
0
5
10
15
20
25
30
35
40
45
50
Figure 7: ATS signal from fiber optical.
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0
20
40
60
80
100
120
Figure 8: FFT spectrum with microphone.
0 1 2 3 4 5 6
x 10
4
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Figure 9: Time signal from microphone.
0 2 4 6 8 10 12
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
Normal
Unnormal
Figure 10: Adaptive Kalman filtering algorithm by using
acoustic energy (the dotted line represents a failure
condition while the solid line represents a normal
condition).
4 CONCLUSIONS
A real time monitoring system for Automobile Tire
Set (ATS) using an acoustic signal was proposed.
The point of Taiwan’s Productivity 4.0 plan is how
to plan prediction technology, smart factory, IoT,
and IoV; distributing predicting sensors and the
function of real-time computing and predicting are
implemented in this paper, also, by using intelligent
LAN devices, one can instantly monitor the ATS
with IPII.
The practical sensing of acoustic and fiber
optical signals real-time computing and the adaptive
order theory successfully predict the system
condition of the ATS before a breakdown, which
helps the follow-up maintenances and alarm
instruction. This paper brought up intelligent ATS
equipment based on fiber optical and acoustic
signals, which can effectively monitor various
operational conditions of ATS. The IPII system uses
acoustic signals to predict the shap, tire depth
features of ATS, so that the losses can be reduced to
the minimum before the vehicle accidents.
The result of the experiment proved that the
adaptive order analysis method was an effective
diagnosis to be applied on intelligent ATS real-time
prediction; however, the amount of feature energy
value shown by each big data order needs to be
distinguished by applying an intelligent ambiguous
system, and the order sampling number of every
condition must be taken into consideration, five
samples at least, to increase the accuracy of the
identification rate.
ACKNOWLEDGEMENTS
This study was supported by the Ministry of Science
VEHITS 2016 - International Conference on Vehicle Technology and Intelligent Transport Systems
172
and Technology of Taiwan, the Republic of China,
under project number MOST 104-2221-E-182-044.
REFERENCES
He, Y., Xu, 2014. Developing Vehicular Data Cloud
Services in the IoT Enviroment. IEEE Transactions on
Industrial Informatics, vol. 10, No. 2, May.
Haykin, 2002. Adaptive Filter Theorey, Prentice Hall.
America, 4
th
edition.
Wu, B., Su, H., 2009. An expert system for the diagnosis
of faults in rotating machinery using adaptive order-
tracking algorithm, Expert Systems with Applications,
vol. 36.
Real Time Monitoring System for Automotive Tire Set using an Acoustic Signal
173