Abstract
Active and semi-active vehicle suspension systems have been used for decades in the
automotive engineering industry and have been applied to passenger vehicles in pursuit of enhanced
occupant comfort, and improved vehicle handling by reducing the tyre load variations
induced by road irregularities. A lesser-known use of these suspensions is yaw rate control.
This thesis builds on earlier research on the use of active suspension systems for this purpose,
and focuses on implementing Nonlinear Model Predictive Control (NMPC) to capitalise on
the capabilities of these suspension systems.
This study addresses several gaps in the current literature with regards to vehicle state
estimation using Unscented Kalman Filters (UKFs) with adaptive covariance matrices. There
are also several gaps regarding real-time capable NMPC controllers that target several
objective functions simultaneously. The overall aim of this thesis is to develop and enhance
vehicle control using active suspensions and NMPC strategies. This was achieved by
initially developing an enhanced vehicle state estimator, followed by a multifunctional NMPC
controller and subsequently a multifunctional integrated controller that combines an NMPC
controller with other types of controllers. In the first part of the study, the objective (1) was
to improve vehicle state estimation performance in extreme driving conditions by developing
a novel state estimator using a UKF with adaptive covariance matrices. The estimator was
tested across a wide variety of high slip scenarios, in both high and low friction conditions.
The results show that the proposed estimator provides improved accuracy and robustness in
all test cases considered when compared to a well-tuned baseline UKF estimator.
In the second part of the study, the objective (2) was to enhance current NMPC controller
capabilities in relation to vehicle dynamics control, body motion control and active safety, by
developing real-time implementable suspension controllers using the NMPC approach. Two
NMPC controllers - one with adaptable cost function weights - governing both semi-active
and active suspensions, were initially developed and tested against different performance
objectives. The NMPC controllers outperformed several benchmarking controllers in every
test considered, with improvements in both body motion control and yaw rate tracking, even
in situations at the limit of handling.
Finally, in the third part of the study, the objective (3) was to develop a novel integrated
control strategy, comprised of NMPC with road profile preview, Skyhook and PID
front-to-total anti-roll moment distribution controllers. Working in tandem, this Ride-Blending
Controller (RBC) is designed to manage both active suspension and front-to-total torque
distribution. It was subsequently assessed against its individual constituent controllers on
an experimentally validated simulation model. The developed RBC was shown to improve
performance in all test cases considered, across all key performance indicators. The indicators
were designed to highlight the multi-faceted ability of the RBC, and thus relate to body
motion, ride comfort, active safety and yaw rate control.
The novel outcomes developed in the presented study provide enhanced vehicle control
solutions and can be directly implemented in the automotive engineering industry.
Furthermore, the contribution to knowledge and understanding can be further expanded upon in
academia, towards the improved use of active suspension systems and implementation of
NMPC controllers.