Research On Vehicle Dynamics Integrated Control For A Full Drive-by-Wire Vehicle With Four-Wheel-Distributed Steering And Four-Wheel-Distributed Traction/Braking Systems

With the growing maturity of the vehicular X-by-wire technologies, the full drive-by-wire vehicle with four-wheel-distributed steering/traction/braking systems is considered to be a critical part of the future vehicle industry. In the meanwhile, severe traffic security issues nowadays have been leading to an upsurge in demand for active safety technologies. With its unique chassis structure, a full drive-by-wire electric vehicle is the perfect benchmark platform used to evaluate various active safety systems. Hence, it is very necessary to develop the chassis control system specifically for this kind of new-type vehicle, which could theoretically achieve the optimal dynamic responses.Funded by Open Fund Project of the State Key Laboratory of Automotive Simulation and Control ’Integrated Control Method for a Full Drive-by-Wire Electric Vehicle Based on Driver’s Intention Recognition’ (No.20120111), State High-Tech Development Plan (the 863 program) ’Development of the Dynamics Control Systems of Electric Vehicles’(No. 2012AA110904) and Graduate Innovation Fund of Jilin University ’State Estimation and Road Friction Coefficient Identification for the Full Drive-by-Wire Electric Vehicle’(No. 20121088), this dissertation starts from the fundamental principle of vehicle dynamics and studies the hierarchical integrated control method for the distributed steering/traction/braking system, so that the vehicle can achieve the optimal dynamic responses in the aspect of handling, stability and tracking ability. Based on the rapid prototyping technique, a measurement and control platform is built for the full drive-by-wire electric vehicle UFEV, whereby the proposed observer and controllers are validated through low-speed field tests and model-in-the-loop experiments. The main work of this dissertation is as follows:1) On account of the unique chassis structure of the distributed steering/traction/braking system, the rapid control prototyping technique is employed to build the vehicle control unit for the full drive-by-wire electric vehicle, where six time-critical loops are configured and executed simultaneously in order to supervise, control and coordinate the on-board sensors and actuators, as well as to record their signals, so that the proposed observation and control algorithms can be tested on this real-time controller. The basic driving modes at low speed (i.e. front wheel steering, four wheel steering, oblique driving and zero radius turning) are determined based on the kinematic principle, while the finite state machine is utilized to design the transition logic of these modes. The field running tests show that the vehicle control unit meets the development requirement of a measurement and control platform and lets the vehicle achieve omni-directional moving at low speed.2) Based on the dual unscented Kalman filtering theory, a dynamics observer is developed to obtain the vehicle states of the plane motion, which are used by the following integrated chassis controller. This algorithm also identifies the total torque of the in-wheel motors (IWM) on either the left or the right side of the vehicle, such that the observation accuracy and the control effect will be improved. Both the off-line simulation and the field test results show that the state estimator can effectively estimate the longitudinal and lateral velocities of the full drive-by-wire vehicle in the front/four wheel steering modes and filters out the signal noise of the gyro sensor, while the parameter identification module observes the motor torques with good accuracy.3) With its flexible chassis layout and high control degrees-of-freedom (DOF), the full drive-by-wire electric vehicle is an ideal platform for the integrated control. This dissertation presents a model-based hierarchical control method consisting of 5 layers to coordinate the distributed steering/traction/braking system, which places the lateral stability first and the handling performance second. In the driver control layer, a 2-DOF reference model and the optimal preview acceleration model are utilized for the control task, which consider and realize the driver’s anticipation of the vehicle’s handling characteristics. In the motion control layer, A novel MIMO non-linear sliding mode controller calculates the motion ;’control efforts to follow the driver’s commands, whereby the non-singular TSMC technique is employed for the yaw motion control. In the tire force allocation layer, the friction circle is linearized by the proposed polygonal constraints by taking into account of the load transfer due to both roll and pitch. The tire force is distributed by a hybrid strategy with the algorithm divided into no more than three quadratic programming optimizations, such that the total workload of four wheels is minimized during normal driving while the motion control efforts are maximized in extreme handling conditions. In the executive layer, the actuator controller controls all the actuators (i.e. servo steering motors and IWMs) in order to eventually achieve the desired tire forces. Both the CarSim/MATLAB co-simulation and the field test results validate the effectiveness of the integrated controller, which can greatly improve the vehicle’s stability as well as its responsiveness.4) In order to let the full drive-by-wire vehicle achieve the optimal tracking ability and follow a given trajectory at varying speed, the nonlinear model predictive control (MPC) is employed to redesign both the driver control layer and the motion control layer, whereby the spatial transformation is utilized for explicit acquisition of the reference trajectory in the prediction horizon. A hierarchical structure is introduced to improve the real-time performance of the spatial MPC, where the terminal sliding mode control technique is employed to get the state errors and position errors to precisely converge in finite time. By building a driving simulator, the path following algorithm is evaluated through model-in-the-loop simulation, and the results show that the vehicle can best follow a given trajectory at varying speed and achieve the optimal tracking ability with this algorithm.The main innovation points of this dissertation are as follows:1) The presented vehicle control unit considers the unique chassis structure of the full drive-by-wire vehicle and coordinates the whole steering/traction/braking system, so that advanced observation and control algorithms can be tested on this platform. In contrast, only one function can be implemented by a traditional electronic control unit. The basic driving modes are designed based on the kinematic principle, hence the vehicle can achieve omni-directional moving at low speed.2) Most chassis controllers need to precisely know the vehicle’s states of motion, the dynamics observer based on the dual unscented Kalman filter is utilized to filter out the signal noise of the gyro sensor as well as to estimate the longitudinal and lateral velocities. Moreover, this observer also identifies the nominal torque coefficients of the IWMs on the left and right sides of the vehicle, so that the optimal performance of the observer and the integrated controller is assured.3) Based on the sliding mode control method and the optimal distribution theory, this dissertation establishes a complete integrated control system for the distributed steering/traction/braking system. The quadratic programming algorithm with linear equalities and inequalities is employed to fully explore the potential of the vehicle stability. The actuator control module is introduced to eventually ensure the control precision and performance of the integrated controller. The terminal sliding mode control technique is used to further enhance the dynamic response of the yaw motion.4) For the trajectory tracking problem of the full drive-by-wire vehicle, the spatial transformation and the nonlinear model predictive control theory are used to develop the combined longitudinal and lateral motion control algorithm, whereby the hierarchical structure is introduced to greatly improve its real-time performance and the terminal sliding mode control technique gets the state errors and position errors to precisely converge in finite time. Therefore, the motion control task is to finally achieve the optimality.