Caizhen received his Bachelor and Master's degrees in Vehicle Engineering at Jilin University, China. Then he worked as Project Engineer at DaimlerChrysler SIM Technology Co., Ltd, Shanghai, China for two years. After that, he studied his PhD degree at the Cambridge Vehicle Dynamics Consortium (CVDC), University of Cambridge on the heavy vehicle dynamics and control from 2004 to 2008. Then he worked as Development Engineer in Ricardo on the development of hybrid vehicle supervisory control. Currently, he is working as a Research Fellow at the University of Warwick on a hybrid vehicle project of Sustainable Action on Vehicle Energy (SAVE).
Abstract: This paper presents a comprehensive and flexible forward dynamic powertrain simulation tool, WARwick Powertrain Simulation Tool for ARchitectures 2 (WARPSTAR2), for modelling of conventional internal combustion engine, hybrid and pure electric vehicles. WARPSTAR2 includes physical powertrain component models and their controllers, hybrid supervisory controller, the driver and environment model. The physical powertrain component models are developed in Dymola, whilst the component controllers, the hybrid supervisory controllers, and the driver model are realised in MATLAB/Simulink. Thus the powers of these two softwares are combined.
<br><br>
A generalised fuzzy-logic-based supervisory controller is proposed for all hybrid electric vehicle (HEV) architectures so that all HEVs with different architectures share the same structure of supervisory controller. The generalised formation can be used for the supervisory controllers of different HEV architectures with varied parameter settings thus facilitating the controller design process. The rule-based supervisory controller is also developed in WARPSTAR2. Simulation is carried out for different HEVs with these two supervisory controllers in the driving cycles. The results of engine and battery power usages with these two supervisory controllers are similar and the differences of predicted engine fuel consumptions between the two supervisory controllers are within 5%.
Abstract: The real world fuel economy of vehicles is becoming increasingly important to manufacturers and customers. One of the major influences on this is driver behaviour, but it is difficult to study in a controlled and repeatable manner.
<br><br>
An assessment of driver models for studying real world driver behaviour has been carried out. It has been found that none of the currently existing driver models have sufficient fidelity for studying the effects of real world driver behaviour on individual vehicle fuel economy. A decision making process has been proposed which allows a driver model with a range of driving tasks to be developed. This paper reports the initial results of a driver model as applied to the conceptually straightforward scenario of high speed cruising. Data for the driver model has been obtained through real world data logging.
<br><br>
It has been shown that the simulation driver model can provide a good representation of real world driving behaviour in terms of vehicle speed and this is compared to a number of logged driver speed traces. A comparison of modelled fuel economy for logged and driver model real world drivers shows good agreement.
Abstract: A high-speed optimal trailer steering controller for a tractor-semitrailer is discussed. A linear model of a tractor-semitrailer with steered trailer axles is described, and an optimal trailer steering controller is introduced. A path following controller is derived to minimize the path tracking error in steady-state manoeuvres using active trailer steering. A roll stability controller is introduced by adding the lateral acceleration of trailer Centre of Gravity as another objective in the steering controller, so as to improve roll stability in transient manoeuvres. A strategy to switch between these two control modes is demonstrated. Simulation results show that the steering controller can ensure good path tracking of articulated vehicles in steady-state manoeuvres, and improve roll stability significantly in transient manoeuvres, while maintaining the path tracking deviation within an acceptable range. Tests with an experimental tractor-semitrailer equipped with a high-bandwidth active steering system validate the controller design and simulation results. The roll stability controller reduces the measured rearward amplification by 27%.
Abstract: This article discusses algorithms to estimate parameters and states of articulated heavy vehicles. First, 3- and 5-degrees-of-freedom linear vehicle models of a tractor semitrailer are presented. Vehicle parameter estimation methods based on the dual extended Kalman filter and state estimation based on the Kalman filter are presented. A program of experimental tests on an instrumental heavy goods vehicle is described. Simulation and experimental results showed that the algorithms generate accurate estimates of vehicle parameters and states under most circumstances.
Abstract: This paper discusses an optimal linear quadratic control algorithm to improve the roll stability of a tractor semi-trailer using active semi-trailer steering. The controller minimises a combination of the path-tracking deviation of the trailer rear end relative to the path of the hitch point (5th wheel) and the lateral acceleration of trailer centre of gravity (CoG). First a linear vehicle model of tractor semi-trailer is constructed. Then a 'virtual driver' model for trailer steering control is introduced to minimise the path-tracking deviation of trailer rear end. The lateral acceleration of trailer CoG is included as a second objective of the optimal controller so as to improve roll stability. A Kalman filter with linear vehicle model is used to estimate unknown vehicle states, needed by the controller. Simulation results show that optimal control of semi-trailer steering could improve the roll stability significantly during transient manoeuvres while keeping the path-tracking deviation of trailer rear end within an acceptable range.
Abstract: A modelling structure for different architectures of Hybrid Electric Vehicles (HEVs) is presented in this paper. This structure includes physical powertrain components and their controllers, hybrid supervisory controller, and the driver model. The physical powertrain component models are developed in Dymola, whilst the component controllers, hybrid supervisory controllers, and the driver model are developed in MATLAB/Simulink.
<br><br>
The structure makes it possible to compare different hybrid vehicle architectures directly with the capability to study the influence of real-world driver behaviour on energy usage. Three types of HEV architectures, including Mild Parallel, Series, and Power Split, are illustrated to show the feasibility of this modelling technique.
Abstract: A model to investigate the effects of driver behaviour on hybrid vehicle control is presented in this paper. A
simulation model has been developed which can be used to simulate different driver models and thus predict the fuel
consumption for different hybrid vehicle architectures. A cruising study was carried out and the logged data of
acceleration pedal position, gear selection, and road gradient were fed into the simulation model. Two drivers that behave
quite differently from each other in terms of acceleration pedal position are investigated. The fuel consumption figures of
two drivers are compared with a mild parallel hybrid vehicle with rule-based and fuzzy-logic-based supervisory
controllers in simulation. The results indicate that smoother pedal position adjustment in cruising driving could reduce
the fuel consumption and show less switching between operation points of the hybrid vehicle.
Abstract: This paper describes the effects of torsional stiffness in the vehicle drivetrain on 4WD vehicle drivability during transient drive events. The approach uses acausal modelling methods to create a low fidelity torsional stiffness model for a 4WD vehicle during tip-in/tip-out events in response to torque inputs.
<br><br>
The torsional response during these events is one of the key elements to assess the vehicle driveability. The paper focuses on the torsional frequencies in the range of 2Hz to 10Hz, where the human body is very sensitive to oscillation modes. The low fidelity model is correlated with a full nonlinear model. Real time simulation is included in the experiment to validate the response of the torsional model in real time test.
Abstract: An independent rear wheel torque control is developed for a Through-the-Road (TTR) hybrid electric vehicle (HEV) with one electric motor embedded in each of the rear wheels. An 8-Degree-of-Freedom (DOF) nonlinear vehicle model of a TTR HEV is developed, validated and used as plant model in simulation. A simple control strategy with yaw gain derived from this 8-DOF vehicle model is designed and simulated with the plant vehicle model. The performances of the controlled and uncontrolled vehicles have been compared in simulation to investigate the effectiveness of the proposed controllers. The simulation results indicate that the handling behaviour of a TTR HEV with independent rear wheel torque control improves compared to the uncontrolled vehicle.
Abstract: A generalised fuzzy-logic-based power management strategy for various hybrid electric vehicle (HEV) powertrain architectures is presented in this paper. Various HEV powertrain architectures were developed and compared in terms of fuel consumption. For each HEV powertrain architecture, an individual power management controller needs to be designed to manage the power efficiencies of all hybrid components.
<br><br>
In order to facilitate the design process for the power management controllers of various HEVs, a generalised fuzzy-logic-based power management strategy is proposed, which can be easily adapted between different architectures. Then this strategy is implemented in different HEV architectures and simulated in the speed profile driving cycles. The simulation results show that the generalised fuzzy-logic-based power management strategy works well with different HEV powertrain architectures in terms of satisfying the driver’s demand and improving the fuel efficiency.
Abstract: <br>This dissertation describes research on active trailer steering to improve roll stability and safety of articulated heavy vehicles at high speeds, and increase the manoeuvrability and reduce the road and tyre wear at low speeds.
<br><br>
A prototype active semitrailer steering system was available on an instrumental test vehicle and a path following controller was developed in a previous project. The primary aim of this project is to improve roll stability of articulated vehicles at high speeds.
<br><br>
Chapter 1 presents a literature review concerned with the background to trailer steering, rollover prevention, vehicle parameter and state estimation, and integrated steering and braking.
<br><br>
Chapter 2 presents simulation results of various brake actuation events on articulated vehicles with path following control. It is shown that the vehicle is stable under braking with path following control provided the Anti-lock Braking System is functioning. Active trailer steering can compensate the changing direction of trailer body caused by the braking forces and thus maintain the vehicle’s yaw stability.
<br><br>
Chapter 3 and 4 present a strategy for estimating vehicle parameters and states. This strategy is examined in simulation and then validated in vehicle tests. The results show that the algorithms work effectively to estimate the vehicle parameters and states.
<br><br>
Chapter 5 develops an optimal active trailer steering controller at high speeds and a PID controller at low speeds. The high-speed controller ensures good path tracking in steady-state manoeuvres and reduces the lateral acceleration of the trailer CoG in transient manoeuvres. The low-speed controller ensures good path tracking for all manoeuvres. Simulations show that the active trailer steering controller can reduce lateral acceleration significantly, while maintaining acceptable path error.
<br><br>
Chapter 6 presents experimental tests on active steering controller in a full sized test vehicle. The tests verify the simulation results and show that the active steering is beneficial at all speeds. The active steering vehicle improves the roll stability by reducing the lateral acceleration of trailer CoG up to 27% while maintaining path tracking performance as good as a passive vehicle at high speeds. It is more manoeuvrable at low speeds, reducing the tyre wear and causing less road damage.
Abstract: Researchers at the University of Warwick are undertaking a three-year research project to investigate the impact of real world driver behaviour on CO2 emissions and energy use of vehicles. It’s the first time a single modelling framework will be developed for all hybrid vehicle architectures which will link individual vehicle models with driver models, transport models and city electrical energy models with the aim of enhancing understanding and promoting the reduction of energy usage in the transport sector.
Abstract: The development of a driver model for assessing the real world fuel economy of vehicles in a controlled and repeatable manner will be presented. The model incorporates a decision making process that represents a range of driving tasks, and has been validated against multiple data sets gathered from a vehicle operating in cruising and urban driving scenarios.
Abstract: The development of vehicle models for use in assessing driveability and handling behaviour will be described. Models of the powertrain and chassis systems of a four wheel drive vehicle will be presented. Both models have been developed in the object orientated system modelling tool Dymola/Modelica.
