The estimation and motion control of ground vehicles rely on the combination of sensor, actuator and algorithm designs. With the development in all these aspects, vehicle dynamic performance has been increasing over the recent years, but there are still some challenges. One of the major difficulties in vehicle dynamics and control is the ability to estimate the essential parameters for control design. As the environment and vehicle configuration may be subject to change in daily usage, it is desired to have the parameter estimation conducted simultaneously with the control implementation because this will reflect the parameter changes in a very timely manner. However, the integration of estimation and motion control is challenging because a practical vehicle system has sensor noise, model uncertainty and nonlinearities. With electric vehicle powertrain, especially the promising four wheel independent driven (4WID) architecture, some of the above mentioned issues can be alleviated because the electric motor torque can be accurately controlled.
In this dissertation, three topics will be discussed towards the same integrated estimation and motion control goal. The first topic only studies the parameter estimation of vehicle’s center of gravity (CG) position and inertial parameters, that will serve as the basis for building vehicle models. The parameter estimation is based on the Ackermann’s steering geometry (ASG) which benefits the estimation by eliminating the unknown lateral tire force terms in the regression formulation. The vehicle suspension kinematic model is also considered which enables the identification of CG three dimensional positions. Also the proposed method can give the CG three dimensional position simultaneously with vehicle mass and yaw moment of inertia estimation. The second topic will discuss the identification of nonlinear tire force curves through an adaptive control formulation in lateral motion control. The proposed method will extend the conventional adaptive control design to estimate nonlinear mappings. In addition, internal auxiliary variables are used to replace noisy lateral speed and yaw rate signals for smoother control output generation. The third topic will present new algorithms for vehicle braking and traction control applications. The algorithm presented does not assume any conventional tire force model in the braking or traction control, but update the estimated nonlinear curve estimation almost instantaneously during the control process. The control law to stabilize the wheel dynamic system in the descending tire force region is also included in the discussion. The developed estimation and control methods are all validated in the Simulink-CarSim® simulation environment. The braking control performance has also been tested in a chassis dyno platform. The proposed method in this work will enhance vehicle’s performance through estimations about vehicle parameters and tire force models, which may contribute to safety improvements in future vehicle control designs.