IOP Conference Series: Materials Science and Engineering

PAPER • OPEN ACCESS

A comparison of optimal semi-active suspension systems regarding

vehicle ride comfort
To cite this article: Dimitrios Koulocheris et al 2017 IOP Conf. Ser.: Mater. Sci. Eng. 252 012022

View the article online for updates and enhancements.

This content was downloaded from IP address 172.245.152.109 on 24/10/2017 at 05:09

CAR-2017 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 252 (2017) 012022 doi:10.1088/1757-899X/252/1/012022
1234567890

A comparison of optimal semi-active suspension systems

regarding vehicle ride comfort

Dimitrios Koulocheris, 1Georgios Papaioannou, Emmanouil Chrysos

Vehicles Laboratory, School of Mechanical Engineering, National Technical
University of Athens, 15780 Athens, Greece

1
[email protected]

Abstract. The aim of this work is to present a comparison of the main semi active suspension
systems used in a passenger car, after having optimized the suspension systems of the vehicle
model in respect with ride comfort and road holding. Thus, a half car model, equipped with
controllable dampers, along with a seat and a driver was implemented. Semi-active suspensions
have received a lot of attention since they seem to provide the best compromise between cost
(energy consumption, actuators/sensors hardware) and performance in comparison with active
and passive suspensions. In this work, the semi active suspension systems studied are comfort
oriented and consist of (a) the two version of Skyhook control (two states skyhook and skyhook
linear approximation damper), (b) the acceleration driven damper (ADD), (c) the power driven
damper (PDD), (d) the combination of Skyhook and ADD (Mixed Skyhook-ADD) and (e) the
combination of the two with the use of a sensor. The half car model equipped with the above
suspension systems was excited by a road bump, and was optimized using genetic algorithms
(GA) in respect with ride comfort and road holding. This study aims to highlight how the
optimization of the vehicle model could lead to the best compromise between ride comfort and
road holding, overcoming their well-known trade-off. The optimum results were compared with
important performance metrics regarding the vehicle’s dynamic behaviour in general.

1. Introduction
The primary purpose of a vehicle’s suspension system is to provide stability, support, and comfort during
the drive. Ride comfort and isolation from road disturbances are greatly influenced by the suspension
system. Therefore, suspension design is a trade-off between the three conflicting criteria of road holding,
load carrying and passenger comfort. In order to study the suspension system in vehicles, various
mathematical models have been developed such as the quarter car, the half car and the full car model.
The quarter model is dominant in the research studies due to its simplicity [1], however the half car
model demonstrates more accurate results including the effect of either the pitch or the roll phenomena
[2, 3]. Although, for greater accuracy the full car model is used including both the roll and the pitch
phenomena, but it is more demanding computationally than the others and hence very few studies have
been performed based on it. In addition to the vehicle model in recent studies, a seat [4] and a driver [5]
have been added, so as to study the ride comfort of the passengers extensively and investigate in depth
the effects of the vibrations induced by the road to the passengers.
The main focus of these models is the suspension system modeled with a spring and a damper.
Vehicle suspensions are mainly classified into three types depending on their operation: passive, semi-
active and active. Passive suspensions are older conventional systems having non-controlled springs and

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
CAR-2017 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 252 (2017) 012022 doi:10.1088/1757-899X/252/1/012022
1234567890

shock absorbing dampers with fixed parameters. On the contrary, active suspensions can adapt to the
system variations based on online changes of the actuating force. Therefore, active suspensions have
been extensively studied since 1960s and various approaches have been proposed in order to exploit
their advantages However, active suspensions normally require important amounts of power supply,
which is their main drawback, preventing them from being used extensively in practice. From 1970s,
semi-active suspensions have received much more attention, since they can achieve greater performance
than passive ones and consume less power than active ones. Thus, the dominant line of research lies in
the semi-active suspensions since they provide the best trade-off between cost (energy-consumption,
actuators/sensors hardware) and performance.
The research in this field follows two mainstreams: the study of new technologies of semi-active
actuation of damping (like electro-hydraulic, electro-rheological and magnetorheological damper), and
the design of semi-active control strategies. Bourmistrova et al. (2005) [6] applied evolutionary
algorithms to the optimization of the control system parameters of a quarter car model. The multi
objective fitness function which was a weighted sum of car body rate-of-change of acceleration and
suspension travel was minimized. Poussot-Vassal et al. (2010) [9] aimed at providing a full analysis of
the optimal performances both for the comfort and for the handling objectives. These optimal
performances were approximated through a numerical optimization approach, using a Model Predictive
Control model. Lajqi et al. (2012) [1] introduced a design and optimization procedure for active and
semi-active non-linear suspension systems regarding terrain vehicles using a quarter car model. The
excitation of the model was represented by double cosine road bumps. Savaresi et al. (2007) [10]
presented a study of the performance limits of a semi-active suspension system mainly focused on the
objective of comfort. The complementary characteristics of skyhook and acceleration driven damping
(ADD) were combined and the mixed control of the two control laws was investigated. Poussot-Vassal
et al. (2011) [11] evaluated the most classic semi-active control laws in time and frequency domain
using a single corner model. Colina et al. (2014) [12] also investigated the basic semi-active control
laws compared to the passive ones. The model was a quarter car and the general Sine Sweep test
(ISO7401) was used as excitation for the models.
The objective of this paper is to optimize the parameters of passive and semi-active suspension
systems using the Pareto Front optimization method in respect with ride comfort and road holding. The
results are compared using performance metrics. The organization of this paper is as follows: in Chapter
2 a brief description of the models used for both the passive and semi-active cases is provided, in Chapter
3 the road excitation applied is illustrated, while in Chapter 4 the optimization procedure is outlined, its
results are highlighted in Chapter 5 and, finally, in Chapter 6 conclusions are presented.

2. Mathematical models

2.1. Half car model

A half car model, equipped with a seat and a 7-DOF human-body model is considered for analysis. The
model includes the front and rear axle allowing the observation of the pitch phenomena. As shown in
Figure 1, the vehicle model consists of three basic subsystems: the tires, the suspension systems and the
body of the vehicle. The latter is considered as a rigid body of mass m, equal to the half of the total mass
of the vehicle. The distance of the front and rear unsprung masses (mF and mR) from the center of mass
is equal to aF and aR, respectively. The front and rear tires are modelled as linear springs, which receive
as input the irregularities of the road profile. The governing equations of the model have been described
in detail in previous work [2] and the parameters with their values are show in Tables 1 and 2.
Additionally, the seat and the driver of a vehicle are modelled, as shown in Figure 2, in order to
investigate in depth the ride comfort of the passenger. Different parts of the human body such as the
pelvis, the diaphragm, the thorax etc. are described via several m – c – k subsystems. Τhe equations of
the model as well as the parameters were retrieved from the literature [5], [2].
One of the most important parameters of a vehicle model is the suspension travel, which is described
as zST,F =zs -zF -aFθ and zST,R =z  z  a θ for the front and rear suspension respectively, and it is used

2
CAR-2017 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 252 (2017) 012022 doi:10.1088/1757-899X/252/1/012022
1234567890

extensively in the control laws of semi active suspension systems as part of their condition values.
Moreover, another parameter crucial for the dynamic behavior of the vehicle is the tire deflection,
described aszT,F =zF -zRoadF and zT,R =zR -zRoadR for the front and rear tire, respectively.

Figure 1. Half Car Model. Figure 2. Seat – Driver Model.

Table 1. Nomenclature of vehicle parameters.

Parameters Subscripts
z Vertical motion coordinate S Body
θ Pitch motion coordinate F Front
zroad Road excitation R Rear
m Mass T Tire
C Damper’s coefficient ST Suspension
K Spring’s stiffness coefficient deflection
a Distance of the centre of the mass of the vehicle
Table 2. Vehicle design parameters.
Parameter Value Parameter Value
aF [m] 0.91 aR [m] 1.55
mF [kg] 25 mR [kg] 25
mS [kg] 520 KTF , KTR [N/m] 2.0·105
Ι [kg  ] 473 e [m] 0.4

2.2. Semi active control laws

The different semi active controls applied to the suspension systems of the vehicle model are illustrated
in this section. All of them are comfort oriented, mostly improving the comfort of the passengers by
adjusting the damping coefficient of the suspension system according to the dynamic behavior of the
vehicle and its response to the applied excitation. The same control law will be applied in each case in
both the front and rear suspension system. Taking into consideration the results of previous studies, most
of the control laws applied in this work were expanded from the ones proposed to the literature so as to
minimize the changes in the damping factor and thus, avoid the energy cost because of unnecessary
changes of the damper’s state. In order to achieve this, a small positive value of the condition used in
each control law, for both the front and rear suspension system (AF and AR), was set as the limit for the
different states, leading the damper’s coefficient to change not only regarding the sign of its condition
but also based on this value.

2.2.1. Skyhook two state damper control (SH-2). The 2-states Skyhook control is an on/off strategy that
switches between soft and stiff damping coefficient. This control law in literature consists of two states
in which the damping factor Ci of the damper (i.e. its fluid viscosity, air resistance, etc.) changes
according to the sign of the product of z ST, i and z S (equation 1).

3
CAR-2017 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 252 (2017) 012022 doi:10.1088/1757-899X/252/1/012022
1234567890

Cmin, i , if z ST, i z S ≤ Ai
Ci =
Cmin, i , if z ST, i z S  -Ai i=F, R (1)
Cmax, i , if z ST, i z S Ai
where z ST,F ,z ST,R are the stroke velocities of the front and rear, respectively, and Cmin,i and Cmax,i are the
minimal and maximal damping factors achievable by the considered controlled damper.

2.2.2. Skyhook linear approximation damper control (SH-L). An improved version of Skyhook control
has been used to achieve variable damping for additional energy saving. The linear approximation of
the Skyhook control includes the change of the damping factor Ci according to the sign of the product
of z ST, iand z S :
C if z ST, i z S ≤ Ai
 min, i
Ci = C min, i if z ST, i z S  -Ai i=F, R (2)
 aC z
max, i ST, i +1-aCmax z S
if z ST, i z S > Ai
 z ST, i
where a ∈ [0, 1] is a tuning parameter that modifies the closed-loop performances and enables the
controller to adjust its stiff value (Cmax, i) according to the needs of the application, saving more energy.
More specifically, when a = 1, this control law is equivalent to the two-state Skyhook control. The
difference relies on the fact that, according to the third expression (when z ST, i z S >A), such a control
provides an infinite number of damping coefficients, providing the ability of tuning it to be continuous.

2.2.3. Power driven damper control (PDD). The power driven damper (PDD) is a semi-active control
law which consists in changing the damping coefficient Ci as:
Cmin, i ,  z ST, i zST, i +Cmin, i z ST, i ≤ Ai
 Cmin, i ,  z def, i zdef, i +Cmin, i z ST, i > Ai
Cmax, i ,  z def, i zdef, i +Cmax, i z ST, i < -Ai
Ci =Cmin, i +Cmax, i
, zST, i ≠0 and z ST, i =0
i=F, R (3)
 2
 -zST, i , otherwise
 z ST, i
This control law depends on the value of the suspension travel and the stroke velocity. The damping
coefficient may be the minimum, maximum, their mean value or the product of the spring stiffness with
suspension travel divided by the stroke velocity.

2.2.4. Acceleration driven damper control (ADD). The ADD control approach is described by the
equation:
Cmin, i if z ST, i z  ≤ Ai
Ci =
Cmin, i if z ST, i z  ≤ -Ai i=F, R (4)
Cmax, i if z ST, i z  > Ai
Thus, this strategy is shown to be optimal in the sense that it minimizes the vertical body acceleration
(z  ). In literature, it is stated that the “switching dynamic” may influence the closed-loop performances.
The swift changing of values of the sprung acceleration leads to quick changes of the damping factor
and this dynamic change creates a noise.

2.2.5. Mixed Skyhook-Acceleration driven damper control (SH-ADD). SH and ADD present
complementary behaviors in terms of performance, with SH providing the best performance at low
frequency (around the body resonance), and ADD ensuring optimality at mid and high frequency
(beyond the body resonance). Therefore, a mixed control law has been proposed in the past using a very
simple but effective frequency range selector  z 2 -a2 z 2 . According to the sign of the main condition,
either the ADD strategy is selected (mid and high frequency dynamics), or the SH strategy is used. The
equation describing this control law is:

4
CAR-2017 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 252 (2017) 012022 doi:10.1088/1757-899X/252/1/012022
1234567890

Cmax, i if  z 2 -a2 z 2 ≤0 and z ST, i z  >0 or [ z 2 -a2 z 2 >0 and z ST, i z  >0]
Ci =  (5)
Cmin, i otherwise
The parameter a represents the frequency limit between the low and the high frequency ranges.
Specifically, the value of a is set at the crossover frequency (in rad/s) between SH and ADD. This law
is almost optimal as it provides a mix of the best performance of the SH and ADD.

2.2.6. Mixed Skyhook-Acceleration driven damper control with 1 sensor (SH-ADD-1s). The current law
is similar to the mixed Skyhook-ADD, being based upon the complementary performances of soft and
hard passive suspensions. Hard suspension is able to damp optimally the body resonance, but without a
desiring filtering at high frequency. On the contrary, a soft suspension ensures the best filtering but with
the drawback of a poorly damped body resonance. This complementarity resembles the complementarity
between the SH and the ADD, described before, however the main difference is that at mid frequency
the passive performance is far from the one achievable by the best possible semi-active suspension. The
law is described by the following equation:
Cmin, i , if z 2 -a2 z 2 ≤ Ai
Ci =  Cmin, i , if z 2 -a2 z 2 ≥Ai i=F, R (6)
Cmax, i , if z 2 -a2 z 2 <-Ai
According to the current value of z 2 -a2 z 2 , the soft
damping condition is selected when it is positive,
otherwise the hard-damping condition is used. Similarly to the mixed SH-ADD, the amount z 2 -a2 z 2 can
be considered as a simple “frequency-range selector”, where the parameter a represents the limit
between the ranges of low and high frequency.

3. Road excitation
In order to study the dynamic behavior of the vehicle an excitation for the model is required. In this
study, a sinusoidal function was used as excitation representing a road bump, as it is shown in Figure 3.
The geometrical aspects of the bump were chosen as follows: its height was set to 0.05m and its length
was set to 5m. The velocity of the vehicle was considered constant, at 8 m/s. Finally, a time lag between
the front and rear wheel of the vehicle was applied. Specifically, the two wheels follow the same
a +a
trajectory with a time delay due to the distance aF +aR between them (tdistance = FV R) [3].

Figure 3. Road Bump.

4. Optimization Procedure
A multi objective optimization was applied to the model in respect to ride comfort and road holding
using the Pareto front for several constraints. As design variables of the optimization, the controllable
dampers and the spring stiffness were selected, as shown in equation 7:
design variables=[Cmin, F ; Cmax, F ; Cmin, R ; Cmax, R ; KF; KR ] (7)
where Cmin, F , Cmax, F, Cmin, R and Cmax, R were the minimum and maximum achievable damping
coefficient of the front and rear damper, respectively, according to the semi active control law that is in
use (Equations 1-6). Additionally, in semi active control laws SH-L, SH-ADD and SH-ADD-1s the

5
CAR-2017 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 252 (2017) 012022 doi:10.1088/1757-899X/252/1/012022
1234567890

tuning coefficient a was added as a design variable. The bounds of the design variables are illustrated in
Table 3.
Moreover, the optimization of the passive model was carried out for two different bounds for the
damping coefficient of the front and rear suspension. At first, the bounds of the minimum damping
coefficient of the semi active laws were used (PS-LP) and then the ones of the maximum damping
coefficient were applied to the design variables (PS-UP). The bounds of the springs’ stiffness were the
same. The objective functions of the optimization were selected so as to improve both the ride comfort
and the road holding. Therefore, two targets were chosen: the root mean square of the seat’s acceleration
(f1), representing the ride comfort, and the average of the variances of both the front and rear tire
deflections (f2), representing the road holding (Table 3). The constraints used in the optimization
problem were selected so as to incorporate the practical considerations into the design process, such as
the working space of the suspension system, as well as to enhance the optimization targets (Table 4).
Table 3. Design variables, bounds, objectives and constraints.
Lower Upper
Design Variables Objectives Constraints
Bounds Bounds
Cmin, F , Cmin, R (Ns/m) 500 2500 zST, i ≤ 0.127 m
f1 =rms( z 9 )
Cmax, F , Cmax, F (Ns/m) 2500 5000 zT, i ≤ 0.051 m
KF , KR(N/m) 15000 70000 zi ≤ 0.07 m
1
a (SH-ADD, SH-ADD-1s) f2 = !varzT,F +varzT,R "
10 60 2 max( z 9 ) ≤ 4.5 m/s2
(rad/s)

5. Results
The Pareto fronts of all the aforementioned cases are illustrated, in Figure 4. It is obvious that the Pareto
fronts are located in three different areas. In the first area (A1) lay the Pareto fronts of four cases: SH-
2, SH-L, SH-ADD and PS-UB, in the second one (A2) lays the Pareto front of SH-ADD-1s, while the
Pareto fronts of ADD, PDD and PS-LB lay in the third area (A3) which is dominated by the previous
ones, providing the “less optimal” solutions. The results belonging in A1 reach the lowest values of f1
(objective for ride comfort), whilst, in all the studied cases, regardless area, f2 (objective for road
holding) is in same range, apart from the PS-UB which has achieved the highest road holding of all
cases (minimum value of f2).

Figure 4. Comparison of the optimal solutions of all the cases

As it can be observed, in Figure 4, all the Pareto fronts of the semi active suspension laws are within the
area created by the Pareto Front of PS-UB (black) and PS-LB (cyan), proving the main idea of semi
active suspensions, which is the combination of a stiff and a soft damper. Another issue that should be
mentioned is the difference between the Pareto of the passive case studies (PS) compared to the one of
the semi-active case studies (SA). All the SA Pareto fronts reveal a different behavior during their
attempt to increase the road holding of the vehicle. Instead of following a common trajectory like the
Pareto of PS, which reveals the well-known conflict between ride comfort and road holding (increase of

6
CAR-2017 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 252 (2017) 012022 doi:10.1088/1757-899X/252/1/012022
1234567890

the one leads to the decrease of the other), they are demonstrating a vertical “jump off” for f2 values
(circled areas in Figure 4) and then, the Pareto front becomes almost parallel to the axis of f1. This
behavior illustrates the inability of the SA cases to further improve the road holding after a certain point,
which is acceptable considering the fact that the applied control laws are comfort oriented.
The optimal design variables delivered from the Pareto front are illustrated in Figure 5. For each
control law (Figures 5a-5f), eight curves of C-K are plotted. The four of them illustrate the optimal
solutions of cases PS-LB and PS-UB (CF-KF and CR-KR for each case), while the other four demonstrate
the optimal solutions of the SA cases (CFmin-KF, CFmax-KF, CRmin-KR and CRmax-KR). Each plot is divided
in two areas with a red line and in the left one CFmin-KF and CRmin-KR are compared with the results of
PS-LB, whilst in the right one CFmax-KF and CRmax-KR with the results of PS-UB. A “jump off” similar
to the one of the Pareto fronts is again obvious. All the curves concerning the front suspension system
of SA cases, shape an “L” regarding the values of C and K. This L-shape means that the optimization
has concluded in various values of K for a specific value of C (vertical part of the L-shape), as well as
in various values of C for a specific value of K (the horizontal part of the L-shape). The fact that the L-
shape appears in the front suspension system, validates the necessity of using different suspension
systems to the front and rear axles. On the other hand, in all the SA cases, the optimal solutions of the
rear suspensions system have an almost fixed value for the spring stiffness and many alternatives for
CRmin and CRmax, indicating a convergence to the optimization procedure.

(a) (b) (c)

(d) (e) (f)

Figure 5 Design variables C-K figures
Table 4. Performance metrics
# Metric SH-2 SH-L ADD PDD SH-ADD SH-ADD-1s Pass lb Pass ub
1 RMS( z # )[m/s2] 0.293 0.318 0.359 0.352 0.301 0.334 0.373 0.311
2 zST, F [m] 0.019 0.019 0.023 0.026 0.019 0.017 0.024 0.015
3 zST, R [m] 0.019 0.020 0.017 0.023 0.022 0.011 0.021 0.013
4 VDVhead 5.727 6.001 6.342 6.487 5.637 6.077 6.545 6.085
5 Fhead
CF 3.691 3.715 3.544 3.580 3.607 3.540 3.456 3.340
6 zT, F [m] 0.004 0.004 0.005 0.005 0.004 0.004 0.004 0.003
7 zT, R [m] 0.002 0.002 0.003 0.003 0.003 0.002 0.003 0.002

Finally, some important metrics for the vehicle dynamics are illustrated in Table 4. The metrics
calculated here are the root mean square value of the seat’s acceleration, the front and rear suspension

7
CAR-2017 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 252 (2017) 012022 doi:10.1088/1757-899X/252/1/012022
1234567890

travel, the vibration dose value (VDV) and the crest factor of the head [2] and the front and rear tire
deflection. The values included in Table 4 are the mean values of the optimal solutions in order to present
the mean dynamic characteristics of all the solutions delivered from the Pareto front. The current
comparison validates the conclusions stated previously. ADD and PDD do not deliver “good optimal”
results, being unable to improve both the passengers’ ride comfort (line 1 of Table 4) or the road holding
of the vehicle (lines 6 and 7 of Table 4). Whereas, SH-2, SH-ADD and SH-ADD-1s have successfully
converged to solutions either equal or better than the ones of PS-UB, which is the main reference case.

6. Conclusions
To sum up, the main aim of the current work was to present optimized semi-active suspensions along
with optimized passive ones in order to compare them both in terms of the optimization results and the
dynamic characteristics of the solutions. Moreover, in this work a different approach was proposed
regarding the conditions used in the formulas of the control laws, so as to minimize the unnecessary
changes in the damper’s state and save energy for the system. Additionally, during the multi-objective
optimization applied to the problem, remarks were illustrated regarding the well-known conflict of the
suspension system and the attempt of the semi-active suspension systems to overcome it. Future work
is in progress not only to expand this work for semi active control laws which are road holding oriented
but also to implement decision algorithms to the Pareto solutions, so as to figure out the importance of
the “jump” mentioned in the previous sections.

References
[1] Lajqi S and Pehan S 2012 Designs and optimizations of active and semi-active non linear
suspension systems for a terrain vehicle Strojniški vestnik-Journal of Mechanical
Engineering 58.12 732-43
[2] Koulocheris D, Papaioannou G and Christodoulou D, 2017 “Optimal design solution among
pareto alternatives for vehicle nonlinear suspension system”, Proceedings of 26th International
Automotive Conference “Science and Motor Vehicles” (26th JUMV), Beograd, Serbia, 2017.
[3] Koulocheris D V, Papaioannou G D and Christodoulou D 2016 Assessment of the optimization
procedure for the nonlinear suspension system of a heavy vehicle Mobility and Vehicle
Mechanics 42 17-35
[4] Shirahatt A, Prasad P S S, Panzade P and Kulkarni M M 2008 Optimal design of passenger car
suspension for ride and road holding Journal of the Brazilian Society of Mechanical Sciences
and Engineering 30.1 66-76
[5] Abbas W, Emam A, Badran S, Shebl M and Abouelatta O 2013 Optimal seat and suspension
design for a half-car with driver model using genetic algorithm Intelligent Control and
Automation 4.02 199
[6] Bourmistrova A, Storey I and Subic A 2005 December Multiobjective optimisation of active and
semi-active suspension systems with application of evolutionary algorithm In The Proceedings
of the 2005 International Conference on Modelling and Simulation Melbourne Australia
[7] Poussot-Vassal C, Savaresi S M, Spelta C, Sename O and Dugard L 2010 A methodology for
optimal semi-active suspension systems performance evaluation In Decision and Control
(CDC) 2010 49th IEEE Conference on pp 2892-97 IEEE
[8] Savaresi S M and Spelta C 2007 Mixed sky-hook and ADD: Approaching the filtering limits of a
semi-active suspension Journal of dynamic systems, measurement, and control 129.4 382-92
[9] Poussot-Vassal C, Spelta C, Sename O, Savaresi S M and Dugard L 2011 Survey on some
automotive semi-active suspension control methods: A comparative study on a single-corner
model IFAC Proceedings Volumes 44.1 1802-07
[10] Colina A, Lerma G, Cabanes I and Iglesias I 2014 Modelling and control of a semi-active
suspension system In New Advances in Mechanisms, Transmissions and Applications pp 25-
32 Springer Netherlands