1

CHAPTER 1

INTRODUCTION

1.1 GENERAL

Suspension system of an automobile plays a vital role in vehicle

handling, maneuverability, performance and ride comfort. Vehicle handling
depends on the force acting between the road surface and the wheels. Ride
comfort is related to vehicle motion sensed by the passenger. To control ride
comfort and vehicle handling, various types of suspension systems are used
namely, passive, semi active and active suspension system.

Most of the automobile industries have been using passive

suspension system for decades and it consists of coil spring or a leaf spring in
parallel with a viscous damper. Semi-active suspension system consists of
passive elements like spring and variable damper. It can only change the
viscous damping coefficient of the damper and does not add energy to the
suspension system.

Active suspension system consists of a separate force actuator

which exerts an independent force on the suspension system to improve the
ride comfort and vehicle handling ability. The primary requirement of the
suspension system is that it should minimize the vertical acceleration in the
body to ensure passengers ride comfort, minimize the dynamic tyre load to
provide maximum road handling while cornering and braking.
 2

1.2 PASSIVE SUSPENSION SYSTEM

A passive suspension system is a conventional suspension system

that consists of a spring and shock-absorbing damper whose characteristics
are fixed. The commercial vehicles today use passive suspension system as
means to control the dynamics of a vehicle’s vertical motion as well as pitch
and roll. Passive suspension system controls the motion of the vehicle body
and wheel by limiting the suspension velocity according to the rate
determined by the designer but they do not provide energy to the system.
Hence, the performance of a passive suspension system is variable subject to
the road profiles.

1.3 ACTIVE SUSPENSION SYSTEM

Active suspension is one in which the passive components are

augmented by an actuator that supply additional forces into the system. These
additional forces are determined by the control law using the data from the
sensors attached to the vehicle. The actuator is normally attached in parallel
with both the spring and shock absorber. Since the hydraulic actuator
connects the unsprung mass to the body, it can control both the wheel hop
motion as well as the body motion. Thus, the active suspension can improve
both the ride comfort and road handling capability simultaneously. The main
advantage of an active suspension is the associative adaptation potential
where the suspension characteristics can be adjusted while driving to match
with the profile of the road being traversed. Terminologies related to
suspension performance are explained as follows.

Ride Comfort – is related to vehicle body motion sensed by

the passengers. By bringing down the sprung mass
acceleration, it can be improved.
 3

Suspension Deflection – is a relative displacement between

the vehicle body (sprung mass) and the wheel (unsprung
mass). It is constrained within an allowable range of motion.

Road Handling – is associated with the contact forces between

the road surface and the vehicle tyre. The relative displacement
between the unsprung mass and the road represents the tyre
deflection.

Moreover active suspensions use actuators, sensors and control

algorithms to introduce bi-directional control forces to the suspension system.

1.4 LITERATURE REVIEW

Active suspension systems have been widely studied over the last
three decades. Various control strategies such as optimal state-feedback, back
stepping method, fuzzy control have been proposed in the past to control the
active suspension system (Hrovart (1993)). This section presents a brief
review on the summary of reported work on optimal controller, Observer
design, Fuzzy Logic Controller (FLC), Sliding Mode control (SMC), Fuzzy
Sliding Mode Control (FSMC), Force tracking and Active force control of an
automotive active suspension system.

1.4.1 Optimal Control

Optimal control is based on Linear Quadratic (LQ) control theory

where a vehicle model is assumed to be a linear or approximate linear system.

Thompson (1976) and Wilson et al (1986) applied optimal control

theory to the design of an active suspension system. The performance index
used is based on ride quality, suspension deflection and tyre deflection.
 4

Tseng and Hrovat (1990) addressed the main characteristics of optimal active
suspension based on two degree-of-freedom quarter-car vehicle models and
associated global performance.

Engelman and Rizzoni (1993) designed an active suspension

system and reported that Linear Quadratic Gaussian (LQG) controller fails to
achieve the desired performance since the actuator force had to be kept small
to prevent instability.

Narayanan and Senthil (1998) presented an active control of the

time varying response to a stationary random excitation of a two degree of
freedom vehicle model. Equivalent linearization method is used to derive the
vehicle model and optimal control laws are obtained by using stochastic
optimal control theory based on full state information. The performance of an
active suspension with nonlinear passive elements is found to be superior to
the corresponding passive suspension system.

Senthil Kumar and Vijayarangan (2006) developed an active

suspension system for a quarter car by designing a controller using LQ
optimal control theory, based on acceleration dependent method and achieved
34.08% reduction in passenger acceleration compared to the passive one.

Zhang et al (2007) presented a feed forward and feedback

controller design approach to solve the problem of optimal control for vehicle
active suspension systems with control delay.

Kashtiban et al (2009) and Iraj et al (2010) developed a nonlinear

optimal control law based on a quadratic cost function and applied on a half-
car model for the control of active suspension systems.
 5

1.4.2 Observer Design

Rajamani and Hedrick (1995) developed an adaptive observer for

identification of the sprung mass of the automobile. A realistic model of the
suspension system incorporating the dynamics of the hydraulic actuator is
discussed. The observer is used to adapt on dry friction present in hydraulic
actuator and on spring stiffness, viscous damping and hydraulic bulk
modulus.

Yoshimura et al (2001) developed Variable Structure System (VSS)

observer for the active suspension system. SMC is used as a control scheme
and the active control force is obtained by using a pneumatic actuator. Road
profile is estimated by using a simplified VSS observer. The proposed active
suspension system shows an improvement in vibration isolation of the car
body than LQ control.

Adizul et al (2005) implemented Proportional Integral Sliding

Mode Control (PISMC) strategy with observer design for an active
suspension system. Simulation results demonstrate that PISMC with
disturbance observer gives better performance than Linear Quadratic
Regulator (LQR) controller. The road profile is estimated by using an
observer. Using Lyapunov stability theory, it is shown that the estimation
error is bounded stable.

Dixit and Buckner (2005) implemented Sliding Mode Observation

and Control (SMOC) strategy to a semi active vehicle suspension system
using model reference approach. Sliding Mode Observer (SMO) is developed
to estimate the unmeasured state variables that require only suspension
deflection as a measured input. The performance and robustness of SMC,
SMO and SMOC are demonstrated through comprehensive simulations.
 6

Simulation results demonstrate the effectiveness of SMOC in improving the

ride quality.

Cheng et al (2010) designed a full-car active suspension controller

with a reduced order observer for vehicles so as to improve ride comfort and
reduce the suspension deflection. Three kinds of road profiles, a bump road, a
random white noise and a power spectral density road profile are used in the
study. Simulations demonstrate that the proposed FSMC can provide the best
ride comfort and the least suspension deflection under all these road profiles.

1.4.3 Fuzzy Logic Controller

Ting et al (1995) developed a fuzzy control scheme for active

suspension control system on the basis of reduced order model analysis and
SMC principle. Fuzzy control rules are constructed on self tuning
methodology instead of human expertise. Set of rules are built-in as the look-
up table of microcomputer for practical implementation. The robustness of the
controller is also discerned. Chattering phenomenon inherent from the SMC is
eliminated due to the fuzziness of the FLC. Simulation results show
satisfactory performance under perturbed conditions.

Barr and Ray (1996) designed and simulated a FLC for an active
suspension system. The performance is compared with LQG controlled active
and passive suspension. Neither LQG nor FLC scheme showed an
improvement in road handling. But ride comfort is superior in FLC based
active suspension control.

Yoshimura (1996) developed an active suspension system using

fuzzy logic. The active control is determined such that the passenger ride
comfort is minimized under the constraint of the suspension deflection and
tyre deflection. The stability of the proposed active suspension system is
 7

obtained by tuning the control rules and membership functions. Simulation

results show that the proposed active control is more effective than passive
control in terms of the ride comfort of passengers.

Rao and Prahlad (1997) proposed a novel FLC based active

suspension system. The parameters in FLC like, the spread, flex and the
centre of bell-shaped membership functions are fine tuned by trial and error
method. Simulation studies demonstrate the effectiveness of the FLC based
active suspension systems in bringing down the suspension deflection and
acceleration to a level to that of a hypothetical reference model. Also the
performance of the FLC under perturbed condition is studied.

Kuo et al (1999) proposed a Genetic Algorithm (GA) based fuzzy

PI/PID controller for an active suspension system. GA is used to optimize
fuzzy PI/PD controllers to determine the rule table. Simulation results
demonstrate that the proposed controller achieves better performance
compared with passive and linear optimal control.

Huang and Chao (2000) proposed a FLC theory with grey predictor
for an active suspension system to improve the ride comfort. A model-free
fuzzy control algorithm is employed for achieving vibration isolation. The
experimental results show that the tyre deformation influences significantly
the control performance of the active suspension system. Hence, a grey
predictor is introduced to predict the tyre deformation and filter it from the
feedback error signal. The control performance of this fuzzy control strategy
with grey predictor is significantly improved.

Liu et al (2004) proposed a fuzzy control for active suspension

model. High speed on/off solenoid valves are used as control valves.
Experimental results show that the active suspension system can achieve
 8

better vibration isolation performance than passive suspension system, the

displacement amplitude of automobile body is reduced to 55%.

Cao et al (2008) proposed a novel adaptive fuzzy logic controller

based on an interval type-2 fuzzy controller for vehicle non-linear active
suspension systems. Lin and Lian (2011) propose a Hybrid Self-organizing
Fuzzy and Radial Basis-function Neural-network Controller (HSFRBNC) for
an active suspension system. The HSFRBNC uses a radial basis-function
neural-network to regulate the parameters of the Self-Organizing Fuzzy
Controller (SOFC) in real time.

1.4.4 Sliding Mode Control

SMC concept is relatively simple in structure of the controller than

fuzzy and neural network and it guarantees the system stability. Yoshimura
et al (2001) proposed an active suspension system for a quarter car model
using the concept of SMC. Road profile is estimated using minimum order
observer based on a linear system transformed from the exact non-linear
system. Experimental results show the effectiveness of the proposed active
suspension system in vibration isolation of the car body.

Nizar et al (2002) presented a robust intelligent nonlinear controller

for an active suspension system based on a comprehensive and realistic
nonlinear model. A sliding mode neural network inference fuzzy logic
controller is designed for automotive suspension system to enhance the ride
comfort. Simulations are performed and the results show that the proposed
control outperforms existing conventional controllers with respect to body
acceleration suspension deflection and tyre deflection.

Orestes and Miguel (2002) applied static and dynamic variable

structure control for active isolation of vibrations in mechanical systems. A
 9

mechanical nonlinear system of one degree of freedom with an electro

hydraulic actuator is taken for study. Simulation results prove the robust
performance of the controlled system in the presence of external disturbances.

Sam et al (2004) proposed a Proportional Integral Sliding Mode

Control (PISMC), a robust strategy in designing a controller for an active
suspension system which is based on variable structure control technique.
Simulation results show that PISMC proved to be effective in controlling
vehicle and more robust compared to LQR method and the passive suspension
system.

Sam and Osman (2006) presented a new mathematical model for a

hydraulically actuated suspension system for the quarter car model. To
achieve the desired ride comfort and road handling and to solve the
mismatched condition, a PISMC technique is presented to deal with the
system and uncertainties.

1.4.5 Fuzzy Sliding Mode Control

Huang and Lin (2003) proposed a novel Adaptive Fuzzy Sliding

Mode Controller (AFSMC) to control an active hydraulic suspension system.
This control strategy establishes the fuzzy rules by continuous online learning
instead of trial and error process. It simplifies the implementation of FLC.
The experimental results showed that this AFSMC had significantly
suppressed the sprung mass position oscillation amplitude. In addition, the
control voltage is smooth and converging speed of the adaptation algorithm is
quick.

Huang and Chen (2006) developed a functional approximation

based adaptive sliding controller with fuzzy compensation to control quarter
car hydraulic active suspension. Fourier series functions are used to
 10

approximate the nonlinear time varying function for designing a SMC. In

addition a fuzzy scheme with learning ability is used to compensate the
functional approximation error. The advantage of this approach is to design a
model free sliding mode controller. The stability of the proposed controller is
guaranteed by means of Lyapunov theorem. Experimental results show that
the proposed control scheme effectively suppresses the oscillation amplitude
of the vehicle sprung mass corresponding to road surface variation and
external uncertainties.

Zeinali and Khajepour (2010) presented a novel adaptive fuzzy

sliding mode control scheme and its application to the height control of laser
metal deposition process. To eliminate the chattering in SMC, the control
scheme proposed to use an adaptive fuzzy inverse dynamic model of the
metal detection process constructed from the input-output data and an
auxiliary continuous proportional-integral type control law constructed using
the algebraic distance of trajectories from the switching surface. Simulation
and experimental studies are presented to verify the performance and
robustness of the controller and its ability in eliminating the chattering and
tracking the different trajectories.

1.4.6 Force Tracking and Active Force Control

Lin and Kanellakopoulos (1997) proposed a novel nonlinear control

design for active suspensions, which uses a filter whose bandwidth is a
nonlinear function of suspension travel. This filter results in an improved
tradeoff between ride comfort and rattle space usage. An adaptive road
algorithm which continuously monitors the suspension travel and adjusts on-
line the shape of the filter nonlinearity is also used. This adaptive design
realizes the true potential of active suspensions, to yield better ride comfort
over wide range of road surfaces than passive suspension.
 11

Chantranuwathana and Peng (1999) presented experimental results

of a force tracking controller for a quarter car active suspension system.
Active suspension controller had two loops. Main loop utilize LQ control for
the force computation. Second loop use an adaptive robust control strategy to
design the force tracking controller in the presence of plant uncertainties.
Simulation results show the effectiveness of force tracking up to 5 Hz.

Anakwa et al (2002) developed a prototype pneumatic active

suspension system. Mathematical model of the suspension system is derived
analytically and validated experimentally. An electronic subsystem is used to
process the sensor signals which are sent to a Motorola 68HC16
microcontroller based evaluation board. The microcontroller controls a four
bit automatic binary regulator that controls airflow to the pneumatic actuator
for damping. The microcontroller is able to reject a physically generated
0.1143-m negative step road disturbance input. The details of the design
construction, modeling, analysis, computer simulation, controller
implementation and experimental results are also presented.

Chantranuwathana and Peng (2004) developed a Modular Adaptive

Robust Control (MARC) technique and applied to design the force loop
controller of an electro hydraulic active suspension system. Experimental
results show that the proposed MARC performs force tracking up to 10 Hz.
Integrated main loop and sub loop performance when investigated the active
suspension outperformed passive suspension in terms of ride quality up to
10 Hz. Tyre deflection performance is improved only up to 2 Hz.

Mehmet et al (2007) proposed an integrated active suspension

controller for vehicle dynamics emulation. The proposed controller consists of
an active body controller and a force controller, both designed based on
mathematical models derived from physical principles and also validated by
 12

experimental data. Simulation results are validated by experiments on a test

vehicle.

Mailah and Priyandoko (2007) proposed a novel Adaptive Fuzzy

logic and Active Force Control (AFAFC) strategy applied to an active
suspension system incorporating a nonlinear hydraulic actuator. Adaptive
fuzzy technique has been used for computation of the control force and for the
estimation of the mass in the active force control loop and PI control loop for
force tracking of the hydraulic actuator. The performance of the controller is
found to be superior compared to AF based outer loop controller.

Priyandoko et al (2009) designed a new and novel controller

employing the Skyhook Adaptive Neuro Active Force Control (SANAFC)
and implemented for the real-time control of a quarter car active suspension.
Simulation results illustrate the superiority of the scheme over its counterparts
in all aspects related to its riding performance. Experimental results
demonstrated that the active suspension based on the SANAFC controller
outperforms the PID controller and passive suspension in terms of sprung
mass displacement and sprung mass acceleration. But suspension and tyre
deflection performance showed no improvement.

Hassan et al (2010) presented a simulation study involving the

application of an Active Force Control (AFC) strategy to suppress vibration
on the rear handle of a handheld tool. Four types of control schemes are
closely examined and compared involving the classic Proportional-Integral-
Derivative (PID) control scheme, the AFC with crude approximation method,
the AFC with iterative learning method and AFC with Fuzzy Logic (AFCFL)
method. Simulation results show that AFCFL scheme provides the best
performance for the AFC-based controllers.
 13

Literature review reveals that most of the proposed control

strategies tested on simulated models and only few results are available with
experiments on real vehicle. Optimal control has been used by most of the
researchers. Since optimal control is a state feedback control, observer design
is given importance in this work. Fuzzy logic is commonly used in control
strategy and the membership functions have been determined by trial and
error method. Hence a method of tuning FLC parameters is given
importance. SMC and FSMC have been reported in some papers for active
suspension control to ensure robustness. Development of a controller that is
simple in structure, robust and at the same time stability is to be guaranteed.

1.5 OBJECTIVE OF THE RESEARCH WORK

The primary aim of the thesis is to design a controller for vehicle

suspension system to enhance the ride comfort. To begin with, outer loop
controller is designed assuming that actuator produces the desired force. Next,
by considering the actuator dynamics, force tracking of the actuator is to be
ensured. Controller need to be robust in the presence of known/unknown
disturbances and uncertainties. The performance of the designed control
strategies are evaluated in terms of ride comfort and road holding ability.

The main objectives of this thesis are

1. To design and compare the performance of an observer based

optimal controller for an active suspension system with the
Passive one.

2. To optimize the FLC parameters using optimization algorithm

and compare their performance in terms of ride comfort and
road holding ability.
 14

3. To design a controller that has simple structure, guarantees

system stability and handle the chattering problem of SMC
effectively for an active suspension system.

4. To highlight the performance of an intelligent active force

control based active suspension system and analyze the
integrated outer loop and the AFC loop.

5. To implement experimentally a laboratory scaled Active

Vibration Isolation (AVI) set up for a single degree of
freedom mechanical suspension system.

1.6 ORGNIZATION OF THESIS

This thesis is organized into seven chapters.

Chapter 2 presents an optimal controller with SMO design for

quarter car model of an active suspension system.

In Chapter 3, FLC design of an active suspension system is

discussed. GA and Particle Swarm Optimization (PSO) are dealt in detail.
FLC parameter tuning with GA and PSO algorithms are discussed to improve
the ride comfort of the suspension system.

Chapter 4 describes the design of SMC, DFSMC for an active

suspension system. Improving the performance of DFSMC by PSO tuning is
also addressed. Simulation results are discussed to evaluate the controller
performance.

Chapter 5 discusses the Mailah and Priyandoko (2007) method of

intelligent active force control of vehicle suspension system. Design of force
tracking loop, AFC loop and the outer loop for target force computation are
 15

discussed in detail. Power Spectral Density (PSD) analysis of sprung mass

acceleration is presented.

Chapter 6 deals with the practical implementation of feedback

control strategy to a laboratory scaled vibration isolator platform.

Chapter 7 summarizes the review of the work carried out and scope
for future research.