Professional Documents
Culture Documents
Applied Soft Computing: Jimoh O. Pedro, Muhammed Dangor, Olurotimi A. Dahunsi, M. Montaz Ali
Applied Soft Computing: Jimoh O. Pedro, Muhammed Dangor, Olurotimi A. Dahunsi, M. Montaz Ali
a r t i c l e i n f o a b s t r a c t
Article history: The core factors governing the performance of active vehicle suspension systems (AVSS) are the inherent
Received 23 February 2013 trade-offs involving suspension travel, ride comfort, road holding and power consumption. In addition to
Received in revised form 14 May 2014 this, robustness to parameter variations is an essential issue that affects the effectiveness of highly non-
Accepted 16 May 2014
linear electrohydraulic AVSS. Therefore, this paper proposes a nonlinear control approach using dynamic
Available online 27 June 2014
neural network (DNN)-based input–output feedback linearization (FBL) for a quarter-car AVSS. The gains
of the proposed controllers and the weights of the DNNs are selected using particle swarm optimization
Keywords:
(PSO) algorithm while addressing simultaneously the AVSS trade-offs. Robustness and effectiveness of
Feedback linearization
Dynamic neural networks
the proposed controller were demonstrated through simulations.
Particle swarm optimization © 2014 Elsevier B.V. All rights reserved.
Active vehicle suspension systems
Ride comfort
Road holding
Introduction network (NN)-based control have been able to deal with these
issues more effectively [2–10]. In contrast to linear controllers,
Vehicle suspensions are subsystems that aim to improve the nonlinear control methods generate an input that aims to remove
performance of an automobile by isolating the vehicle from road- or significantly reduce the effects of nonlinearities in the system.
induced disturbances, improving passenger ride comfort, and The nonlinear AVSS model presented by Shi et al. [3] incor-
enhancing the road holding performance of the vehicle. However, porated a servo-hydraulic system controlled by a combination of
better ride comfort demands a softer suspension and superior road sliding mode and feedback linearization control methods. This
holding requires a stiffer suspension. The desire to manage these effort and those of Yagiz and Sakman [2] and Chamseddine et al.
trade-offs has led to the development of active vehicle suspension [9] using sliding mode control were plagued with chattering chal-
systems (AVSS) which incorporates an actuator to deal with these lenges.
compromises in real time [1]. Backstepping control has been performed by [4,5,10] for half-car
AVSS are highly nonlinear systems with complex actuator and full-car models respectively. Their solutions showed a signif-
dynamics and need to be designed carefully to manage its sensitiv- icant improvement over the passive vehicle suspension system
ity to parameter variations. Nonlinear AVSS have been successfully (PVSS) and additionally provided an adequate bandwidth where
designed with linear controllers [1]. However, these controllers lack they were able to reject a large range of road disturbances, which
robustness when dealing with variations in vehicle speed, sprung emphasizes the robustness of this control technique.
mass and tyre load. On the contrary, nonlinear control schemes However, these preceding control laws are fundamentally based
such as Fuzzy Logic Control (FLC), backstepping, FBL and neural on the mathematical models which have been chosen to be
either linear or nonlinear. In the case of backstepping, highly
nonlinear and realistic models would require rigorous interlaced
backstepping which may in some cases prove impossible to solve.
∗ Corresponding author. Tel.: +27 117177317; fax: +27 117177049.
Furthermore, the associating zero-dynamics present in real models
E-mail addresses: jimoh.pedro@wits.ac.za, jimoh.olarewaju.pedro@gmail.com
(J.O. Pedro), Muhammed.Dangor@students.wits.ac.za (M. Dangor), may be unstable under certain conditions. Furthermore, the sys-
tundedahunsi@gmail.com (O.A. Dahunsi), montaz.ali@wits.ac.za (M.M. Ali). tems may not be completely understood in reality and this would
http://dx.doi.org/10.1016/j.asoc.2014.05.013
1568-4946/© 2014 Elsevier B.V. All rights reserved.
J.O. Pedro et al. / Applied Soft Computing 24 (2014) 50–62 51
require the use of some model predictive controller to learn the systems: linear AVSS [23], semi-active suspension [22,33], evap-
dynamics of the plant. orators [30,36], continuously stirred tank reactors (CSTR) [29,35]
Intelligent nonlinear control seeks to emulate human logic as and flexible manipulators [37].
well as the brain. They do not require the mathematics of the sys- Yildirim [23] successfully identified a linear AVSS using a recur-
tem to be completely understood whilst developing a control law. rent neural network (RNN) and thereafter carried out network
FLC and NN-based control form part of this set and have been suc- inversion to control the system. He achieved an improvement over
cessfully implemented for AVSS designs to deal with robustness the PVSS. Zapateiro et al. [22] performed recurrent neural network
issues and to better manage AVSS trade-offs [13–25]. (RNN)-based backstepping control on a semi-active suspension
Documented works in the literature have demonstrated that that utilized a magnetorheological (MR) damper. Metered et al. [33]
good AVSS performances can be achieved using FLC [24–26]. Com- carried out network inversion on a MR-based semi-active suspen-
bination of FLC with neural networks in AVSS applications has also sion that was identified with a RNN.
shown remarkable improvements in the robustness of the con- Becerikli et al. [34] presented a DNN to identify and control
troller designed. Rajeswari and Lakshmi [14], Lian [26], and Aldair a CSTR. The system displayed adequate performance in the pres-
and Wang [18] have proposed hybrid neuro-fuzzy controllers. They ence of a wide range of disturbances. The start-up and regulation
realized that the control structure performs better than a NN- problems of CSTR were resolved better with this configuration as
based PID control architecture, which makes it more suitable for compared to the currently employed control schemes for CSTR
AVSS applications. Furthermore, there is no need to mathemati- tanks.
cally model the system since the NN can approximate it through Al Seyab and Cao [35] created a DNN to identify and control
the process of system identification. a double CSTR plant. It was concluded that the DNN decreased
The accomplishments of FLC in AVSS have made it a suitable the training time and improved the accuracy in the identification
candidate for optimization algorithms. Chiou et al. [13], Rajeswari process as compared to conventional model predictive control con-
and Lakshmi [14], and Pekgökgöz et al. [25] used evolutionary algo- figurations. Nanayakkarra et al. [36] successfully trained a DNN to
rithms to derive the membership functions of a FLC. This method identify an evaporator with the use of an evolutionary algorithm.
was successful in improving either the body-heave acceleration The DNN outperformed the static NNs and required fewer neurons
or the suspension travel with larger success than a PID-controlled to learn the dynamics of the plant.
AVSS. Tian and Collins [37] designed a neuro-fuzzy controller for a
With regards to the application of NN-based control to AVSS, flexible manipulator. The dynamics of the nonlinear model was
intelligent controllers using multilayer NNs in system identifica- learnt using a DNN. The identification results were satisfactory and
tion and control have improved the AVSS response as compared to this control method was superior to conventional industrial robots
the PVSS. Tang et al. [15] investigated the performance of a half-car controllers. Deng et al. [28], Deng et al. [29] and Garces et al. [30]
AVSS that was controlled using a multilayer feedforward NN and utilized DNN-based FBL for a variety of control systems. In each
genetic algorithm (GA). There was an improvement in the passen- case, the network was trained using genetic algorithm and the
ger’s seat vertical response as compared to that of a PVSS. In terms response of the system displayed superior results as compared to
of training a multilayer NN through PSO for AVSS, Alfi and Fateh conventional control architectures. Additionally they noticed that
[17] showed that this method performs better than the conven- this control law can be implemented with linear control such as
tional NN training algorithms and GA-based training with quicker PID to improve system response.
convergence speeds, improved accuracy, and had no premature Both linear and nonlinear AVSS controllers have been optimized
convergence problem. using heuristic search methods since they are effective methods
Guclu and Gulez [27], and Aldair and Wang [18] utilized net- in finding global minima. These methods are based on a random
work inversion to control a full-car nonlinear AVSS with actuator search methodology and do not require any function based meth-
dynamics. The NN-based controllers for each case displayed supe- ods to find the minimum. Such methods include PSO and GA. Wai
rior performance as compared to the PVSS. Eski and Yildirim [16] et al. [12], Crews et al. [11], Chiou et al. [13], and Rajeswari and
also used an adaptive multilayer NN to create a robust PID con- Lakshmi [14] optimized the gains of a single loop linear PID-based
troller for a full-car model. The system displayed high identification AVSS using PSO and GA techniques respectively. Pekgökgöz et al.
and tracking capabilities as compared to offline supervised learning [25] and Chiou et al. [13] optimized the membership functions and
algorithms. FLC control parameters using GA and PSO. The fitness function of
Pedro et al. [21] designed a direct adaptive NN-based FBL con- these approaches incorporated sprung mass acceleration and body
troller for nonlinear quarter-car AVSS using radial basis function NN displacement. These optimal policies outperformed the manually
(RBFNN). However, the model did not contain any actuator dynam- tuned AVSS in terms of acceleration and body displacement at the
ics and ignored zero dynamics that may exist as a result of FBL. The expense of the actuation force.
ride comfort and road holding improved as compared to the PVSS The major contribution of the paper is the demonstration of the
and PID-controlled AVSS. Pedro and Dahunsi [20] later utilized a impact of the combination of a PSO-optimized cascaded PID control
multilayer feedforward NN to perform indirect adaptive control of a with dynamic neural network (DNN)-based feedback linearization
servo-hydraulic nonlinear AVSS using FBL. They considered subse- control. The weights of the DNN are chosen using PSO algorithm.
quent zero dynamics and their resulting system displayed superior Performance results for these controllers are presented for each
performance as compared to the case where linear controllers were controller when applied to a 2 degree-of-freedom (DOF) non-
used. linear electrohydraulic vehicle suspension system. These results
DNN uses differential equation to model the neuron and con- are also compared with that of a controller that combines both
tain feedback elements. DNN offers several benefits above static schemes. The controller with combined controlled schemes is char-
neural networks (radial basis function neural network (RBFNN) acterized with better performance for all the vehicle suspension
and multilayer perceptron neural network (MLPNN)) especially as performance parameters considered. The controller is also tested
regard computational efficiency. DNN has capacity to learn com- for robustness using response to parameter variation within the
plex nonlinear systems especially when static neural networks fails 20% range for speed, mass and tyre stiffness. The difference in the
to represent the model appropriately [30–32]. Application of DNN results obtained is marginal. Frequency domain analysis of the con-
with PSO training is very rare, especially with respect to AVSS con- trollers are also presented within the whole-body-vibration range
troller design. However, DNN has been used for various control of 0–80 Hz. The trend in performance is the same, corroborating
52 J.O. Pedro et al. / Applied Soft Computing 24 (2014) 50–62
Both the spring and damping forces are functions of the sus-
the previous results and showing that all controllers attenuated pension travel (xw − xc ) and suspension travel velocity (ẋw − ẋc ),
vibration disturbance within the frequency range considered. respectively. The suspension components have linear, symmetric
The paper is organized as follows. In Active suspension system: and nonlinear elements which are fundamentally a function of the
overview and modelling section we present the active suspension suspension travel and its velocity and are described as follows [20]:
system overview and mathematical modelling. Controller design
section introduces the AVSS performance specifications followed Fs = ksl (xw − xc ) + ksnl (xw − xc )3 , (3)
by a brief description of the PID controller. PSO algorithm is also sym
Fb = bls (ẋw − ẋc ) + bnl
s |ẋw − ẋc |sgn(ẋw − ẋc ) − bs |ẋw − ẋc |, (4)
described in this section. Detailed description of the DNN-based
system identification and FBL control are presented. Simulation where ksl and bls are the linear spring stiffness and linear damping
results and discussion section presents both for deterministic and constant of the suspension system, ksnl and bnl
s are the corresponding
random disturbance inputs. Results for robustness analysis are also nonlinear spring stiffness and damping constant of the suspension
presented and discussed in this section. Finally, concluding remarks sym
system, and bs is the associating symmetric damping constant.
and recommendations for future work are given in Conclusion and The elastic behaviour of the tyre is assumed linear and the force
future work section. produced due to its interaction with the road is:
Fw = kt (w − xw ), (5)
Active suspension system: overview and modelling
where (w − xw ) is the deflection of the tyre.
A schematic of the quarter-car model used in this investigation The actuator force Fa is manipulated through an electrohy-
is presented in Fig. 1. The mass of the wheel assembly is mu and that draulic servo-valve which aims to return the system to rest after
of the chassis is ms . These two components are coupled through the the vehicle is disturbed by the road disturbance. A schematic of the
suspension elements (spring ks and damper bs ). In the case of AVSS, actuator explaining the flow of hydraulic fluid and pressure changes
an actuator is placed in parallel with the suspension elements and in the system is shown in Fig. 2.
supplies an actuator force Fa which supports the PVSS components. The electrohydraulic servo-valve consists of two components: a
The flexural nature of the wheel is captured by means of a spring voltage-regulated electromechanical device and a three-land four-
with stiffness kt . way spool-valve hydraulic system. The dynamics of the actuator
Reference frames are created at the wheel, chassis and road are described through Newtonian fluid mechanics. The governing
surface with xw , xc and w denoting the vertical movement of the equations for the electrohydraulic actuator may be structured to a
wheel, vertical movement of the chassis and the road profile respec- simpler form that is suitable for FBL using the following relations
tively. The associated velocity and accelerations of these bodies are [38,39,41]:
represented as ẋ and ẍ, respectively.
The governing equations of the system are derived by applying P˙L = ˚xv − ˇPL + ˛A(ẋw − ẋc ), (6)
Newton’s laws to both the wheel and chassis:
where ˚ = 1 × 2 with 1 = sgn(Ps − sgn(x
v )PL ) and 2 =
ms ẍc = Fks + Fbs − Fa , (1) (Ps − sgn(xv )PL ), ˛ =
4ˇe
, ˇe = ˛Ct p, = Cd S 1
Vt . A is the cross-
mu ẍw = −Fks − Fbs + Fa + Fw , (2) sectional area of the piston, PL is the change in pressure experienced
across the piston, xv is the servo-valve displacement, Ps is the sup-
where Fks and Fbs are the respective spring and damping forces ply pressure into the hydraulic cylinder, Pr is the return pressure
exerted by the suspension, Fw is the force produced by the from the hydraulic cylinder, Pu and Pl represent the oil pressure
road disturbance and Fa is the actuator force. It is assumed that in the upper and lower portion of the cylinder respectively, Vt
the wheel and suspension elements compress such that w > xw is the total actuator volume, ˇe is the effective bulk modulus of
and xw > xc . the system, ˚ is the hydraulic load flow, Ctp is the total leakage
J.O. Pedro et al. / Applied Soft Computing 24 (2014) 50–62 53
coefficient of the piston, Cd is the discharge coefficient, S is the The disturbance matrix w is represented by:
spool-valve area gradient and is the hydraulic fluid density.
T
In order to reduce complexity, it is assumed that the electrome- w(t)
w(x) = 0 0 0 kt 0 0 . (13)
chanical device that controls the motion of the spool valve is a mu
first-order element with a time constant and is described as:
The elements of these matrices are as follows [20]:
1
ẋv = (Kv u − xv ), (7)
f1 (x) = x3 , (14)
where Kv is the valve gain and u is the control input voltage. Values
of the parameters used in the quarter-car model are given in Table 1. f2 (x) = x4 , (15)
1
sym
f3 (x) = ksl (x2 − x1 ) + ksnl (x2 − x1 )3 + bls (x4 − x3 ) − bs |x4 − x3 | + bnl
s |x4 − x3 |sgn(x4 − x3 ) − Ax5 , (16)
ms
1
sym
f4 (x) = −ksl (x2 − x1 ) − ksnl (x2 − x1 )3 − bls (x4 − x3 ) + bs |x4 − x3 | − bnl
s |x4 − x3 |sgn(x4 − x3 ) + Ax5 , (17)
mu
Fig. 5. Architecture of the DNN-based FBL control augmented with PSO algorithm.
condition. The next condition may only be fulfilled after the selec- • Seed strength of 22,641.
tion of appropriate input–output data. • 0.001 s sampling time.
An important step in system identification is to select a range iii. Hyperbolic Tangent is used for the activation function (x) as
of input–output data that covers the range of signals that will be this ensures the DNN stability [28–30].
encountered in reality. In indirect adaptive control the DNN must
predict the output of the suspension system for a given set of control With regards to pruning and the choice of hidden layer size nn the
input voltage. White-Band-Limited (WBL) noise is used to create a response of the DNN is analyzed for a range of nn starting from
set of input data because it can successfully create a random set of one. The network size was increased until satisfactory results were
input signals which span the space of the expected input signals. attained for a credible range of randomly selected network param-
Selection of an appropriate data set is a rather rigorous process eters. Fig. 7 shows the general trend of the suspension travel output
that requires several conditions to be met. Firstly, the dynamics of for the various hidden layer sizes. It is evident from this figure
the subsystem with the smallest time constant must be captured that a hidden layer size nn of 8 is capable of capturing the sensi-
and this demands that the seed strength of the WBL be significantly tive dynamics of the system as it can pick up the sudden rate of
high. Secondly, the input data must span the space of all possible change of suspension travel more adequately than the 4-neuron
input voltages, which is known to be in the range of ±10 V. Similarly, and 6-neuron configurations. Hence, it is suitable for system iden-
the set of suspension travel output must span the region in which it tification. The structure of the DNN is summarized in Table 2.
is expected to operate, which corresponds to ±0.1 m. “Whole Body This DNN is trained offline using PSO algorithm. PSO was first
Vibration” (WBV) frequency range is classified to be between 0.5 introduced by Kennedy and Eberhart [44] and they mimic the food
and 80 Hz. Human occupants are susceptible to vibrations within searching process of swarms. As in the swarm of animals, each
this range. The resonant frequency range for the suspension sys-
tem is smaller and fits into both the WBV as well as, the “low Table 2
frequency” range. It is thus paramount that the system identifi- Configuration of applied dynamic neural networks for the quarter-car system.
cation input covers these frequencies [42]. In order to satisfy the
Property Numerical value
preceding conditions, WBL was set as follows:
Number of hidden layers 2
Number of neurons in first hidden layer 8
i. Control input u operates within ±10 V. Number of delayed system inputs 1
ii. WBL has the following properties: Number of delayed system outputs 1
56 J.O. Pedro et al. / Applied Soft Computing 24 (2014) 50–62
-3
0.08 10
Actual
0.06 4 neurons
6 neurons
Suspension Travel (m)
0.04 8 neurons
-0.02 -5
10
-0.04
-0.06
0 5 10 15 20 -6
10
Time (s) 0 20 40 60 80 100 120 140 160
Number of Iterations
Fig. 7. Quarter-car suspension travel output for different hidden layer sizes.
Fig. 8. Evolution of the fitness function, (MSE), during the offline PSO-optimized
training of the DNN.
animal or particle searches a search space for a specific item. Each
particle thereafter relays their success in relation to finding the containing the personal best parameters of each particle.
desired item. Each particle will subsequently travel towards the rand1 and rand2 are pseudo-random numbers.
position of the particle with the highest success at various speeds. Step 7: If the fitness of its new position is better than the fitness of
At the same time, it will examine the area for a personal best posi- its personal best, replace its personal best position with its
tion to achieve. The position of the particle refers to the location current position and proceed to Step 8, otherwise continue
of the particle in n dimensional space where n is the number of straight to Step 8 without any adjustments.
parameters that are being altered. Searching towards the vicinity Step 8: Find the fittest particle in the population and choose it as
of the best particle and in the neighbourhood its personal best are the best particle and return to Step 4.
referred to as the global and local searches respectively. The pro-
cess repeats itself until the stopping criterion is met. The algorithm In this learning process, the DNN parameters , W, ␥, and are the
is summarized in the following steps [17]: problem variables that are determined by the PSO algorithm.  is a
1 × 8 vector with ˇn denoting the time constant of the nth neuron
Step 1: Produce a random population of particles using the uniform from the 8 present in the first hidden layer. The same applies to ␥
distribution. and as well. W is the weighting matrix that connects each of the
Step 2: Provide an initial small velocity to each particle using neurons of the hidden layer to each other and it is a square 8 × 8
pseudo-random normal distributions. matrix with Wjh denoting the feedback weighting of the hth neuron
Step 3: Choose the fittest particle as the best particle. into the jth neuron.
Step 4: If the stopping criteria has been met, then stop the algo- The objective function of the PSO algorithm during the system
rithm, otherwise proceed to Step 5. identification process is the mean square error (MSE) of the devia-
Step 5: Adjust the local and global search parameters according to: tion between the actual and predicted outputs [20]:
n
1
N
wlocal = w1 , (28) 2
nmax J = MSE = (yi − ŷi ) , (32)
2N
n
i=1
wglobal = w2 1 − , (29)
nmax where N is the total number of samples used in the input–output
data. The PSO parameters chosen are listed in Table 3 and the
where wlocal , and wglobal are the local and global search
variation of the global fitness with each iteration is presented in
parameters respectively; whilst w1 and w2 are the maxi-
Fig. 8. The resulting verification and validation data are shown in
mum local and global search weighting respectively. n and
Figs. 9 and 10, respectively.
nmax are the current iteration and maximum number of
Before controller design can begin, the stability of the DNN
iterations respectively.
model must be ensured. Deng et al. [28], Deng et al. [29] and Garces
Step 6: Compute the new positions of each particle in the search
et al. [30] suggest that the DNN stability will be guaranteed if the
space using:
followingconditions hold:
X(t + 1) = X(t) + V (t + 1), (30)
Table 3
where X is a matrix comprising of the neural network Parameters used during the PSO implementation to obtain the DNN model.
parameters, V is a matrix consisting of velocity as each
Parameters Value
parameter varies
Population size 200
V (t + 1) = w1 V (t) + wlocal × rand1(Pbest − X(t)) + wglobal Maximum number of iterations 200
Velocity inertia weight (w1 ) 0.5
× rand2(Gbest − X(t)), (31) Maximum social inertia weight (w2 ) 2
Maximum global inertia weight (w3 ) 2
where Gbest denotes the vector containing the parame- Fitness function MSE
Goal for MSE 9 × 10−6
ters of the global best particle, Pbest represents the matrix
J.O. Pedro et al. / Applied Soft Computing 24 (2014) 50–62 57
As the activation function (x) is the hyperbolic tangent func- + . . . + W18 (1 − (x8 )2 )] [f(x) + g(x)u] ,
tion, conditions i. and ii. are fulfilled. With the network computed
through PSO algorithm, both Eq. (33) and the inequality of Eq. = d(x) + e(x)ut = L2f h1 (x) + Lg Lf h1 (x)ut ,
(34) are satisfied. Hence, it may be concluded that the DNN model
attained through training is indeed stable. where d(x) or L2f h1 (x) is the free response of the system and e(x)ut
or Lg Lf h1 (x)ut is the forced response of the system. In the above
derivative of the output, the DNN of the PSO training yielded
a matrix where g1 (x), g2 (x), . . ., g3 (x) were considerably large
constants. Hence the computation of the second derivative of the
0.06 Actual
Predicted network output ŷ¨ produced a solution which was explicitly depend-
ent on the control input ut . Hence, the relative degree of the system
0.04 is two, which infers that the DNN is input–output linearizable as its
Suspension Travel (m)
relative degree is less than the number of states of the DNN (which
0.02 corresponds to 8 as there are eight neurons in the hidden layer).
The next step in the controller formulation demands that the
0 DNN dynamics now be transformed into a coordinate system which
separates the observable and zero dynamics. The DNN may be
-0.02 described in terms of its observed and unobserved zero dynamics
using the diffeomorphism as follows:
-0.04
˙ = f0 (, ), (38)
that:
Controllers Inner loop
2 ŷ¨ = . (43) kp kd ki
u= − 2 L2f ŷ(x) − i−1 Li−1 ŷ(x) . (48) had been traversed. Furthermore, it settled quicker than both the
f
1 Lg L1f ŷ(x) PID + PSO and DNNFBL-controlled cases and this finding suggests
i=0
that DNNFBL + PSO case contains good disturbance rejection prop-
The new virtual control input is determined through a multi-loop erties. These positive results were attained as a consequence of
PID control system described in Fig. 5.
With regards to controller gains, there are now 9 controller gains
1
to be optimized; namely the 6 PID gains of the multi-loop PID con- 10
PID+PSO
troller, and the 3 FBL controller gains (0 , 1 , 2 ). The performance DNNFBL+PSO
index used to select these gains is the same presented in Eq. (20).
The process of the manual tuning of the intelligent controller was
rather cumbersome and rigorous as very small variations in 0 , 1 ,
Performance Index
Table 5
Performance evaluations for the passive, PID-controlled, DNN-based feedback linearization controlled and PSO-augmented, DNN-based feedback linearization controlled
suspension systems to a deterministic road bump.
0.01
Wheel Deflection (m)
0.005
0
0
-0.005
-0.05 -0.01
-0.015
-0.1 -0.02
0 1 2 3 4 5
Time (s) 0 1 2 3 4 5
Time (s)
Fig. 12. Comparison of the suspension travel responses for the PVSS, optimized
PID-controlled, DNN-based feedback linearization controlled and PSO-augmented, Fig. 14. Comparison of the wheel deflection responses for the PVSS, optimized
DNN-based feedback linearization controlled suspension systems to a deterministic PID-controlled, DNN-based FBL-controlled and PSO-augmented, DNN-based FBL-
road bump. controlled AVSS to a deterministic road bump.
60 J.O. Pedro et al. / Applied Soft Computing 24 (2014) 50–62
3000
DNNFBL
PID+PSO
DNNFBL+PSO
2000
1000
Force (N)
-1000
-2000
-3000
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (s)
Fig. 15. Comparison of the actuation force responses for the optimized Fig. 17. Comparison of the suspension travel responses for the PSO-augmented,
PID-controlled, DNN-based FBL-controlled and PSO-augmented, DNN-based FBL- DNN-based FBL-controlled AVSS to −20% parameter variations in vehicle speed,
controlled AVSS to a deterministic road bump. mass and tyre loading.
case displayed a lower degree of chattering in these aspects. These the actuator force will naturally be larger. In conclusion this data
results show that the objective function for the PSO algorithm that argues that actuator dynamics play a significant role in AVSS design
contains each of the AVSS trade-offs is an effective tool for resolv- and that it should not be neglected.
ing design specifications particularly on ride comfort, road holding To test the sensitivity to parameter variations of the
and system transient response. On the other hand, the PID + PSO DNNFBL + PSO controller, several realistic suspension parameters
controlled case performed better than the DNNFBL case and fell were altered. These included tyre stiffness, vehicle mass and speed.
marginally short than the DNNFBL + PSO case in terms of peak and Each parameter was varied by ±20% and the corresponding suspen-
RMS values of the wheel deflection and the body-heave accelera- sion travel responses are plotted in Figs. 17 and 18, respectively.
tion. It did also maintain a greater degree of chattering than the Figs. 17 and 18 indicate that the DNNFBL + PSO case has a satis-
intelligent controller. factory robustness to parameter variations. The system remained
Figs. 15 and 16 indicate that the success of the DNNFBL + PSO Bounded-Input Bounded-Output (BIBO) stable, and the steady-
controller was achieved at the cost of actuation force and voltage. state error increased upon the variation of the various parameters.
This is expected as a control input and hence actuation force is Alteration in vehicle mass contained the most severe steady-state
required to bring down the acceleration considerably into the Less error, followed by the adjustments in vehicle speed and tyre loading
Discomfort range of the ISO standards. However, hydraulic actua- respectively. Its transient response did however remain similar to
tors in general are sensitive to changes in control input, where a the original case where no topology was lost whatsoever. The opti-
slightly smaller variation in control voltage as in the case of the mized PID controller on the other hand displayed no steady-state
optimized PID and DNNFBL controllers produces a larger variation error, but its transient response was worst than the DNNFBL + PSO
in actuator forces. Hence, in order to satisfy both hard design spec- controller with greater settling times and additional peaks. This
ifications (such as ride comfort) and a reduced performance index, implies that the intelligent controller has a better robustness to
parameter variations.
4
DNNFBL
PID+PSO DNNFBL + PSO
3 DNNFBL+PSO 0.085
0.055 Speed Mass Tyre stiffness
2
0.015
1
Voltage (V)
-0.025
0
-0.065
0 1 2 3 4 5
-1 PID + PSO
0.07
0.055
-2
0.015
-3
-0.025
-4
0 1 2 3 4 5 -0.065
Time (s) 0 1 2 3 4 5
Fig. 16. Comparison of the control input voltage responses for the optimized Fig. 18. Comparison of the suspension travel responses for the PSO-augmented,
PID-controlled, DNN-based FBL-controlled and PSO-augmented, DNN-based FBL- DNN-based FBL-controlled AVSS to +20% parameter variations in vehicle speed, mass
controlled AVSS to a deterministic road bump. and tyre loading.
J.O. Pedro et al. / Applied Soft Computing 24 (2014) 50–62 61
0.04 2
10
Passive
0.03 PID+PSO
DNNFBL+PSO
0
0.02 10
0.01
Road Profile (m)
-2
Power Ratio
0 10
-0.01
-4
-0.02 10
-0.03
-6
10
-0.04
-0.05
0 1 2 3 4 5 -8
Time (s) 10 0 1 2
10 10 10
Frequency (Hz)
Fig. 19. Random road profile.
The optimized PID case reported larger peaks with a marginally Parameter Setting
greater RMS value. This implies that the DNNFBL + PSO performed
Computation Algorithm Welch
better in this aspect as well, which further highlights its improved Windowing function Hanning
robustness. Number of points included in fourier transform (NNFT) 1024
System sensitivity is also investigated through frequency Length of window (NWind) 256
domain plots. The most sensitive frequencies of vibrations for Sampling frequency 80 Hz