Hindawi

Complexity
Volume 2020, Article ID 7179801, 8 pages
https://doi.org/10.1155/2020/7179801

Research Article
A Case Study on Computer-Based Analysis of the Stochastic
Stability of Mechanical Structures Driven by White and Colored
Noise: Utilizing Artificial Intelligence Techniques to Design an
Effective Active Suspension System

Aydin Azizi
School of Engineering, Computing and Mathematics, Oxford Brookes University, Wheatley Campus, Oxford No. 33 1HX, UK

Correspondence should be addressed to Aydin Azizi; [email protected]

Received 16 April 2019; Revised 6 November 2019; Accepted 29 January 2020; Published 25 April 2020

Academic Editor: András Rontó

Copyright © 2020 Aydin Azizi. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The goal of this research is to design an Artificial Intelligence controller for the active suspension system of vehicles. The Ring
Probabilistic Logic Neural Network (RPLNN) architecture has been adopted to design the proposed controller, and the pavement
condition has been modelled utilizing Gaussian white noise. The results show that the proposed RPLNN controller has an effective
performance to reduce the unwanted stochastic effect of the road profile.

1. Introduction particle swarm optimization has been adopted by Olivera

et al. [7] to reduce vehicle emissions and fuel consumption
1.1. Background. During the car transportation due to in the city. In 2016, Lotfan et al. [8] combined ANN and
conversion of absorbed energy by suspension and tire of the nondominated sorting genetic algorithm II to model and
vehicle to thermal energy, an energy loss happens. Utilizing reduce CO and NOx emissions from a direct injection
an effective suspension system gives an opportunity to re- dual-fuel engine. ANN has been utilized by Perrotta et al.
duce the energy loss and save energy [1], and saving energy [9] in 2017 to model fuel consumption of trucks. In 2018,
means reducing the fuel consumption. In recent years, many Azizi [10] proposed a PID controller to control fuel
researchers focus on reducing the undesirable effects of the consumption rate based on the changing pavement
pavement and decrease the fuel consumption rate by de- condition. Although the result of the research shows that
signing effective suspension systems [2]. the proposed PID controller can reduce the fuel con-
In 2012, Ismail et al. [3] proposed an Artificial Neural sumption, but the vibration at the beginning of the control
Network (ANN) model for a light-duty diesel engine procedure because of the input control signal for few
powered using blends of various biodiesel fuels with seconds is high.
conventional fossil diesel. Cay et al. [4] utilized ANN to The current research aims to design an effective
predict the brake specific fuel consumption. In 2014, Roy Artificial Intelligence (AI) architecture to decrease the
et al. [5] modelled the performance and emission pa- effect of the road profile which shows itself as imposed
rameters of a single-cylinder four-stroke engine under vibration on vehicle. It means less fuel consumption and
CNG-diesel dual-fuel mode by ANN. In 2015, the per- higher parts life can be achieved by filtering such a noise.
formance, emission, and combustion characteristics of a In this research, Gaussian white noise has been adopted
single-cylinder, four-stroke variable compression ratio to model the road profile. Different from previous works,
engine were predicted with the aid of ANN by Mur- the aim of the current research is presenting an effective
alidharan and Vasudevan [6]. As an intelligent technique, approach to eliminate the mentioned fluctuation and
2 Complexity

control the fuel consumption rate even at the beginning and analyzing the performance of the proposed RPLNN
of the control process [11, 12]. For this purpose, Ring controller and comparing it with the performance of the PID
Probabilistic Logic Neural Network (RPLNN) which is controller.
an effective paradigm and has been utilized by re-
searchers as a function optimizer, prediction tool, and
plant simulator [13–16] is adopted here to be trained by 2.1. Active Suspension Controller Model. The quarter car
the PID controller proposed by Azizi et al. [10]. In this model, which is a model to investigate the dynamic vibration
paper, MATLAB software and Simulink toolbox have of a vehicle with regards to the road profile as the input of the
been utilized to develop a simulation-based model of the model, has been illustrated in Figure 1. It is important to
active suspension system of the vehicle and design an know that, in such a model some parameters such as geo-
effective RPLNN controller to investigate and reduce the metrical information, pitching, and rolling angles have not
effect of the imposed vibration from the road profile. The been neglected, but most of the important features such as
proposed simulator is capable of simulating the vertical the change of the load has been considered.
displacement of the vehicle before and after utilizing the In the current work, to generate the mathematical model,
AI controller, also it is capable of analyzing these vi- the state space method which is one of the well-known
brations by examining the different values of the system approaches in control engineering has been adopted [20].
parameters. The results show that the proposed RPLNN To generate the state space representation, the first step is
controller has a superior performance than the PID analyzing the quarter car model (Figure 1), which has been
controller and has an effective performance to reduce the carried out by the following equations:
imposed vibration by the road profile and reducing the
fuel consumption rate. M1 x€1 (t) + D 􏼂x_ 1 (t) − x_ 2 (t)􏼃 + k1 􏼂x1 (t) − x2 (t)􏼃 � u,
After the brief introduction and literature review (1)
about the research topic, the remaining parts of the paper
is organized in three more sections. First, in Section 2, M2 x€2 (t) − D􏼂x_ 1 (t) − x_ 2 (t)􏼃 + k1 􏼂x2 (t) − x1 (t)􏼃
mathematical modeling of a proposed active suspension (2)
+ k2 􏼂x2 (t) − x0 (t)􏼃 � −u.
of a quarter car model proposed RPLNN algorithm, and
the simulation-based model set up have been discussed in M1 (300 kg) and M2 (40 kg) are the sprung and un-
detail. In Section 3, which is the results section, the sprung masses, respectively. K1 (15000 N/m) and D
designed simulator has been verified by analyzing the (1000 N/m) are stiffness and damping coefficients of spring
effect of the proposed controller’s effect on system vertical and damper of the suspension system, respectively. K2
displacement. Finally, in Section 4, the summary of the (150000 N/m) is the stiffness coefficient of the spring which
research has been given and possible future works have is the model of the tire of the vehicle. The vertical dis-
been suggested. placements of the sprung mass, unsprung mass, and
pavement condition have been shown by X1 (t), X2 (t), and
2. Methodology X0 (t), respectively, and the parameter U represents the
controlled force.
Fuel consumption rate can be affected by two important Now, the state variables with respect to the dynamic
factors: vehicle handling performance and vehicle suspen- equations (1) and (2) and can be defined as follows:
sion system [17]. In conventional method, a passive sus-
pension system which is a combination of springs and x1 � x2 (t),
dampers has been utilized to reduce the imposed vibrations
to the vehicle from the pavement profile. It is important to x2 � x1 (t),
know that, since the damping ratio cannot be adjusted, so (3)
such a system is not able to eliminate the imposed vibration. x3 � x_ 2 (t),
Nowadays, by advancement in technology active suspension x4 � x_ 1 (t).
systems which are equipped with effective controllers are
being utilized to reduce the effect of pavement conditions. By substituting state variables presented on equation (3)
[18]. in equations (2) and (3), we will get the following
The ultimate goal of this research has been defined as to expressions:
design an effective AI controller to reduce the undesirable
effect of the road profile on vehicles. The steps of the pro- D D K K u
x_ 4 � − x4 + x3 − 1 x2 + 1 x1 + , (4)
posed methodology in this research have been defined in the M1 M1 M1 M1 M1
following order. The first step is introducing the quarter car
model equipped with an active suspension system. The D D K K K K u
x_ 3 � x4 − x3 − 1 x1 + 1 x2 − 2 x1 + 2 x0 − ,
second step is utlizing RPLNN algorithm to design an ef- M2 M2 M2 M2 M2 M2 M2
ficient controller\as a part of the active suspension system. (5)
The third step is utilizing MATLAB software to train the
proposed AI controller. The next step is simulating the In general state space variables, change can be defined as
proposed set up utilizing Simulink, and the last step is testing follows:
Complexity 3

Y � CX + DU (10)
Sprung Mass : M1
In this paper, the output has been defined as the dif-
X1 ference of the vertical displacement of the sprung and un-
K1
sprung masses:
Y � x2 (t) − x1 (t). (11)
Control
D U Controller So, based on equation (11) matrixes C and D of equation
signal
(10) will be as follows:
C � 􏼂−1 1 0 0 􏼃,
(12)
D � 􏼂0 0 􏼃.
Unsprung Mass : M2
X2 After introducing the mathematical model of the quarter
car model, the next step is investigation of one of the inputs
K2
of this mathematical model which is the pavement
condition.

X0 2.2. White Noise Road Input Signal. Mathematical modeling

of the pavement condition due to its Gauss-distributed
Figure 1: Quarter car model equipped with the active suspension stochastic behavior is too hard. To fulfill this task, the
system [19]. Gaussian white noise which shows randomness behavior can
be utilized as the representative of the statistical pavement
power spectral density. As shown in equation (13), the re-
dX lationship of the random road input signal and the Gaussian
� AX + BU, (6)
dt white noise per ISO/TC108/SC2N67 can be modelled
mathematically as shown in the following equation [21]:
T
(7) 􏽱�������
X � 􏼂 x 1 x2 x3 x4 􏼃 .
q(s) 2πn 0 Gq n0 􏼁v
� . (13)
So, based on equations (4) and (5) matrixes A and B of w(s) s + 2πf0
equation (6) will be as follows:
where q (s) is the random road input signal, w(t) is the
0 0 1 0
⎢
⎡
⎢ ⎤⎥⎥⎥ Gaussian white noise, n0 is Reference Spatial frequency, Gq
⎢
⎢ ⎥⎥⎥
⎢
⎢
⎢ ⎥⎥ (n0) is the Road roughness coefficient, and f0 is the Filter
⎢
⎢
⎢
⎢
⎢
⎢ 0 0 0 1 ⎥⎥⎥⎥⎥ lower-cut-off frequency.
⎢
⎢
⎢ ⎥⎥⎥ After introducing the quarter car mathematical model
⎢
⎢
⎢
⎢ ⎥⎥⎥⎥
A �⎢⎢
⎢ k + k k D D ⎥⎥⎥, and the Gaussian white noise as one the input signals of it,
⎢
⎢
⎢
⎢ − 1 2 1
− ⎥⎥⎥⎥ the next step is the design of the controller utilizing the
⎢
⎢
⎢ M M M M ⎥⎥⎥
⎢
⎢
⎢ 2 2 2 2 ⎥⎥⎥ RPLNN algorithm.
⎢
⎢
⎢ ⎥⎥⎥
⎢
⎢
⎢ ⎥⎥⎥
⎢ ⎥
⎢
⎢
⎢
⎣ k1 k D D ⎥⎥⎥⎦
− 1 − 2.3. Ring Probabilistic Logic Neural Networks Controller.
M1 M1 M 1 M 1
(8) Probabilistic Logic Neuron (PLN) is a weightless neuron
0 0 which its output can be defined as one, zero, or “don’t care”
⎢
⎡
⎢
⎢ ⎤⎥⎥⎥ states [22]. Because of the “don’t care” state, the PLNs have
⎢
⎢
⎢ ⎥⎥⎥ flexibility to bind together and design a fast-learning
⎢
⎢ ⎥⎥
⎢
⎢
⎢
⎢
⎢ 0 0 ⎥⎥⎥⎥⎥ weightless neural networks [14]. Based on this property the
⎢
⎢
⎢ ⎥⎥⎥
⎢
⎢
⎢ ⎥⎥⎥ RPLNN has been proposed by Menhaj et al. [23] as a novel
⎢
⎢ ⎥⎥⎥
B �⎢⎢
⎢
⎢ k 2 1 ⎥⎥⎥. algorithm to model, control, and optimize complex plants.
⎢
⎢ − ⎥⎥
⎢
⎢
⎢
⎢
⎢ M2 M2 ⎥⎥⎥⎥⎥ The proposed design consists of PLNs in which the output
⎢
⎢
⎢ ⎥⎥⎥ of the last one is fed by the input of the first one (see
⎢
⎢
⎢ ⎥⎥⎥
⎢
⎢
⎢ ⎥⎥⎥ Figure 2).
⎢
⎣ 1 ⎥⎦
0 In 2016, Azizi et al. [15, 22] utilized RPLNN structure as
M1 an optimizer tool to optimize weightless neural networks
and RFID networks. Later, the RPLNN structure has been
The system input variable matrix is
developed and utilized as a part of the hybrid optimizers
U � 􏼂x0 (t) u 􏼃T . (9) [13, 14, 16, 24, 25].
In the current research, different from previous works,
The vehicle suspension system output matrix equation is the weightless RPLNN has been utilized to be trained by the
4 Complexity

O O O O O

1 2 3 . N–1 N

I I I I I
Figure 2: RPLNN structure [13].

u
u y +
fcn – a1 v1 x1
– 1/m1 1/s 1/s
+
+
RPLNN Controller
D +
–

k1

x1–x2
|u|

–
+
– a2 v2 x2
– 1/m2 1/s 1/s
+
–
+ Controlled
output

k1
Pavement
condition
k2

k2

Band-limited
white noise

Figure 3: MATLAB Simulink Model of the Noise Cancellation system for the quarter car model.

proposed PID controller in previous reserach [10], so after objective function the RPLNN controller will be able to
the accomplishment of the training process, the proposed adjust the active suspension force. It is important to know
RPLNN has been utilized as the controller to reduce the that to train the RPLNN controller the reference modelis the
vertical vibration of the vehicle due to the road conditions data of the PID controller, which has been proposed by Azizi
which can be considered as Gaussian white noise input. at el. [10, 26]:
The input of the RPLNN will be the vertical displacement
of the vehicle, and the output will be the actuator's force. It E(t) � x1 (t) − x2 (t). (14)
means that the proposed controller regulates the active
suspension system’s force considering the vibration due to
road condition. 3. Results
To fulfill this task, it is essential to define an objective
function. Since the proposed algorithm gives a fitter result in As seen in Figure 3, the proposed quarter car model has two
each iteration, so it can be called as an evolutionary algo- inputs: the first one is the road condition which is taken as
rithm. Here, since the aim is reducing the difference between random input and has been modelled by Gaussian white
the vertical displacements of the sprung and unsprung noise function, and the second input is the generated force
masses, so the objective function can be defined as the error by the RPLNN controller to reduce the effect of the imposed
function (see equation (14)). In this case by minimizing the noise to the vehicle.
Complexity 5

Pavement condition 1

–1

–2

–3

–4
0 5 10 15 20 25 30 35 40 45 50
Simulation period
Pavement condition
Figure 4: Pavement condition.

3
Uncontrolled displacement

–1

–2

–3

–4

–5
0 5 10 15 20 25 30 35 40 45 50
Simulation period
Uncontrolled displacement
Figure 5: Uncontrolled displacement.

As seen in Figure 4, first input, which is the road con- has been reduced to 0.001.The output of the systemby uti-
dition, has been modelled by Gaussian white noise and the lizing RPLNN controller has been calculated and shown in
effect of this noise, which is the vertical displacement of the Figure 9.
car, has been shown as uncontrolled displacement in Figure 5. By comparing Figures 5 and 9, which are the indicators
As seen in Figure 5, the quarter car model fluctuates of the behavior of the system equipped and not equipped
based on the road input signal. Since it is the uncontrolled with RPLNN controller, it can be observed that the system
behavior, so by comparing Figures 4 and 5, it can be ob- equipped with the RPLNN controller reduces the effect of
served that the vehicle completely follows the road condition the pavement on the vehicle and reduces the vibrations.
which can results in increasing the fuel consumption and Noise cancelation results utilizing PID and RPLNN
compromising passenger’s safety. controllers have been shown in Figures 9 and 10, respec-
As it is mentioned previously, the reference model which tively. By comparing these two figures, it can be observed
has been utilized to train the RPLNN is the data of the PID that both controllers are able to reduce the undesired effect
controller which has been presented in Figures 6 and 7. of the pavement profile on the vehicle. It is important to
The RPLNN controller has been trained offline by mention that although both controllers are able to reduce the
running the training for 100 iterations. As seen in Figure 8, vertical vibrations in the same simulation period, but since
at the end of the training procedure the error function value the active suspension system equipped with the RPLNN
6 Complexity

0.8

0.6

0.4
Controller’s input

0.2

–0.2

–0.4
0 5 10 15 20 25 30 35 40 45 50
Simulation period

Controller’s input
Figure 6: Reference Input data (PID controller) of the controller.

2
Controller’s output

–2

–4

–6

–8

–10
0 5 10 15 20 25 30 35 40 45 50
Simulation period
Controller’s output
Figure 7: Reference output data (PID controller) of the controller.

1.4

1.2

0.8
Error

0.6

0.4

0.2

0
10 20 30 40 50 60 70 80 90 100
Iteration
Error
Figure 8: Error in each training iteration.
Complexity 7

×10–6
1

0.8

0.6
System’s controlled output 0.4

0.2

–0.2

–0.4

–0.6

–0.8

–1
0 5 10 15 20 25 30 35 40 45 50
Simulation period
System’s controlled output (RPLNN)
Figure 9: Noise cancellation utilizing the RPLNN controller.

0.08

0.06
System’s controlled output (PID)

0.04

0.02

–0.02

–0.04
0 5 10 15 20 25 30 35 40 45 50
Simulation period
System’s controlled output (PID)
Figure 10: Noise cancellation utilizing PID controller.

controller has not shown an overshoot response, so it has a vehicle from the road profile (which has been considered as
superior performance than the PID controller. Gaussian white noise), is an effective solution to be con-
To have a better overview on the performance of the sidered as noise cancelation tool, also it is obvious that, due
proposed RPLNN controller, the behavior of the system, to the reduction of vertical vibrations, the lifetime of the car
once equipped with PID and then with RPLNN controllers, parts has been increased.
has been observed and compared.
As seen in Figure 9, at the first stages of the control 4. Conclusion and Future Work
process, the system equipped with PID controller reacts to
the ground excitation with overshoot controlled response, In the current research, to reduce the imposed vibrations to
and after three seconds starts to cancel the noise properly. the quarter car model via the ground excitation an effective
However,as seen in Figure 10, by adopting the RPLNN artificial intelligence controller for the active suspension
controller system from the beginning of the control process system has been designed and modeled. The RPLNN al-
shows a smooth response and prepares a convenient and safe gorithm has been utilized to design the proposed controller,
ride for passengers. and the road input signal has been modelled by utilizing
Finally, it can be concluded that utilizing RPLNN Gaussian white noise. The proposed setup has been modelled
controller, due to decreasing the imposed fluctuations to the by MATLAB and Simulink. The results show that the
8 Complexity

proposed RPLNN controller is able to reduce the undesired Proceedings of the Big Data (Big Data) IEEE International
effect of the road condition on the vehicle, which ultimately Conference on, 2017, IEEE, Boston, MA, USA, pp. 3810–3815,
results in decreasing the fuel consumption and increasing December 2017.
the lifetime of the parts of the vehicle. The performance of [10] A. Azizi, “Computer-based analysis of the stochastic stability
the RPLNN controller has been compared with the PID of mechanical structures driven by white and colored noise,”
Sustainability, vol. 10, no. 10, p. 3419, 2018.
controller and results show that the RPLNN controller has a
[11] A. Azizi and P. G. Yazdi, “Introduction to noise and its
superior performance. As future work, other AI algorithms applications,” in Computer-Based Analysis of the Stochastic
such as Fuzzy Logic and Neural Networks can be utilized as Stability of Mechanical Structures Driven by White and Col-
controllers, also adopting nonlinear models to evaluate the ored Noise, pp. 13–23, Springer, Berlin, Germany, 2019.
performance of the designed controller will be an alternative [12] A. Azizi and P. G. Yazdi, “Noise control techniques,” in
important topic to be investigated. Computer-Based Analysis of the Stochastic Stability of Me-
chanical Structures Driven by White and Colored Noise,
Data Availability pp. 61–73, Springer, Berlin, Germany, 2019.
[13] A. Azizi, “Hybrid artificial intelligence optimization tech-
The related data are presented within the manuscript. nique,” in Applications of Artificial Intelligence Techniques in
Industry 4.0, pp. 27–47, Springer, Berlin, Germany, 2019.
Conflicts of Interest [14] A. Azizi, “Introducing a novel hybrid artificial intelligence
algorithm to optimize network of industrial applications in
The authors declare they have no conflicts of interest. modern manufacturing,” Complexity, vol. 2017, Article ID
8728209, 18 pages, 2017.
[15] A. Azizi, A. Barenji, R. Barenji, and M. Hashemipour,
Acknowledgments “Modeling mechanical properties of FSW thick pure copper
plates and optimizing it utilizing artificial intelligence tech-
This study was funded by the Research Council of Oman
niques,” Sensor Netw Data Commun, vol. 5, no. 142, p. 2, 2016.
under Grant no. ORG/CBS/14/008. [16] A. Azizi, Applications of Artificial Intelligence Techniques in
Industry 4.0, Springer, Berlin, Germany, 2018.
References [17] Y. Zhang, K. Guo, D. Wang, C. Chen, and X. Li, “Energy
conversion mechanism and regenerative potential of vehicle
[1] J. Palmer and S. Sljivar, Vehicle Fuel Consumption Monitor suspensions,” Energy, vol. 119, pp. 961–970, 2017.
and Feedback Systems, Google Patents, Long Beach CA, USA, [18] I. Maciejewski, T. Krzyzynski, and H. Meyer, “Modeling and
2017. vibration control of an active horizontal seat suspension with
[2] A. Azizi and P. G. Yazdi, “Introduction to fuel consumption pneumatic muscles,” Journal of Vibration and Control, vol. 24,
optimization techniques,” in Computer-Based Analysis of the pp. 5938–5950, 2018.
Stochastic Stability of Mechanical Structures Driven by White [19] A. Azizi and P. G. Yazdi, “Mechanical structures: mathe-
and Colored Noise, pp. 1–12, Springer, Berlin, Germany, 2019. matical modeling,” in Computer-Based Analysis of the Sto-
[3] H. M. Ismail, H. K. Ng, C. W. Queck, and S. Gan, “Artificial chastic Stability of Mechanical Structures Driven by White and
neural networks modelling of engine-out responses for a Colored Noise, pp. 37–59, Springer, Berlin, Germany, 2019.
light-duty diesel engine fuelled with biodiesel blends,” Applied [20] D. J. Inman, Vibration with Control, John Wiley & Sons,
Energy, vol. 92, pp. 769–777, 2012. Hoboken, NJ, USA, 2017.
[4] Y. Çay, A. Çiçek, F. Kara, and S. Sağiroğlu, “Prediction of [21] A. Azizi and P. G. Yazdi, “White noise: applications and
engine performance for an alternative fuel using artificial mathematical modeling,” in Computer-Based Analysis of the
neural network,” Applied Thermal Engineering, vol. 37, Stochastic Stability of Mechanical Structures Driven by White
pp. 217–225, 2012. and Colored Noise, pp. 25–36, Springer, Berlin, Germany,
[5] S. Roy, R. Banerjee, A. K. Das, and P. K. Bose, “Development 2019.
of an ANN based system identification tool to estimate the [22] A. Azizi, A. Vatankhah Barenji, and M. Hashmipour, “Op-
performance-emission characteristics of a CRDI assisted timizing radio frequency identification network planning
CNG dual fuel diesel engine,” Journal of Natural Gas Science through ring probabilistic logic neurons,” Advances in Me-
and Engineering, vol. 21, pp. 147–158, 2014. chanical Engineering, vol. 8, no. 8, p. 1687814016663476, 2016.
[6] K. Muralidharan and D. Vasudevan, “Applications of artificial [23] M. B. Menhaj and N. Seifipour, “Function optimization by
neural networks in prediction of performance, emission and RPLNN,” in Proceedings of the 2002 International Joint
combustion characteristics of variable compression ratio Conference on Neural Networks, pp. 1522–1527, Honolulu, HI,
engine fuelled with waste cooking oil biodiesel,” Journal of the USA, May 2002.
Brazilian Society of Mechanical Sciences and Engineering, [24] A. Azizi, “Modern manufacturing,” in Applications of Arti-
vol. 37, no. 3, pp. 915–928, 2015. ficial Intelligence Techniques in Industry 4.0, pp. 7–17,
[7] A. C. Olivera, J. M. Garcı́a-Nieto, and E. Alba, “Reducing Springer, Berlin, Germany, 2019.
vehicle emissions and fuel consumption in the city by using [25] A. Azizi, “RFID network planning,” in Applications of Arti-
particle swarm optimization,” Applied Intelligence, vol. 42, ficial Intelligence Techniques in Industry 4.0, pp. 19–25,
no. 3, pp. 389–405, 2015. Springer, Berlin, Germany, 2019.
[8] S. Lotfan, R. A. Ghiasi, M. Fallah, and M. H. Sadeghi, “ANN- [26] A. Azizi and P. G. Yazdi, “Modeling and control of the effect
based modeling and reducing dual-fuel engine’s challenging of the noise on the mechanical structures,” in Computer-Based
emissions by multi-objective evolutionary algorithm NSGA- Analysis of the Stochastic Stability of Mechanical Structures
II,” Applied Energy, vol. 175, pp. 91–99, 2016. Driven by White and Colored Noise, pp. 75–93, Springer,
[9] F. Perrotta, T. Parry, and L. C. Neves, “Application of machine Berlin, Germany, 2019.
learning for fuel consumption modelling of trucks,” in