| Intelligent Vehicles (IV)which are the integrated carriers of many advanced high-new technologies, are the research frontier of the vehicle engineering and will become a new direction for future development of automobile industry, and their intelligence is mainly incarnated in the automatic navigation. As an important subsystem of the Intelligent Transportation System (ITS), intelligent vehicles have attracted more and more attentions due to the critical traffic safety problems.Navigation control is a key problem in the field of intelligent vehicles research, and it is a research hotspot and the basis of other relevant research. The research contents of navigation control mainly include lateral motion control and longitudinal motion control. Because intelligent vehicles are the nonholonomic motion constraint system, and have highly nonlinear dynamic characteristics and parametric uncertain properties, therefore, how to design the lateral and longitudinal motion control strategy, which can effectively overcome the characteristics of vehicle nonlinear and parameter uncertainties, is the emphasis and difficulty of IV automatic driving and has profound research significance. To solve this problem, in this paper, lateral and longitudinal motion control of intelligent vehicles are studied based on a prototype intelligent vehicle DUTIV-I, which is taken as the platform for experimental research and developed by the School of Automotive Engineering of Dalian University of Technology. The research work of this thesis consists of the following major parts:(1) Image processing algorithms of path information obtained by vision system are designedAn overall design scheme of DUTIV-I’s vision system is descirbed, the selection principle of relevant parameters for CCD camera of DUTIV-I’s vision system is explored, and image processing algorithms of path information obtained by vision system are designed to provide accurate path information for intelligent vehicle navigation control.(2) A vehicle dynamic model is established for simulation research on lateral and longitudinal control systemIn view of the characteristic features of lateral motion control and longitudinal motion control of intelligent vehicle and the structure of the prototype vehicle DUTIV-I, a15degree-of-freedom nonlinear vehicle model which can accurately reflects the vehicle main movement states, the coupling characteristics of the tire model and the nonlinear characteristics of transmission system, is constructed by modularization method based on the basis data of the prototype vehicle. It is composed of engine model, hydraulic torque converter model, automatic transmission model,power train model, brake system model, wheel dynamics model, tire model, vehicle body model, suspension model and actuators model. The verification of established vehicle model is carried out by the simulation and experimental comparison under typical working conditions. The results show that the established vehicle model can accurately describe the vehicle dynamical characteristics and has the high simulation precision, which provides a reliable foundation for researching on lateral and longitudinal control.(3) Study on lateral fuzzy control of intelligent vehicle using genetic algorithms.A vision-based navigation of intelligent vehicle is taken as research object, and the lateral motion control model of vision-based navigation intelligent vehicle is established, which can describe the lateral motion behavior of intelligent vehicles. On this basis, the influence of look-ahead distance and speed on lateral control system is analyzed using root locus method, and the formula of look-ahead distance as a linear function of speed is established, which can optimize the path information data obtained by vision system. Aiming at highly nonlinear and parametric uncertain properties of non-holonomic intelligent vehicles, an automatic lateral controller consisting of a fuzzy feedback control law which can imitate human driving behavior and a feed-forward control law used to offset the disturbance of the curvature of reference paths is designed. In view of the membership functions and rules base designed either by experts knowledge, or iteratively by trial-and error are difficult to adaptive adjust according to the system characteristics and easy to produce overshoot or/and steady-state error. To confront this difficulty, the lateral fuzzy feedback control strategy of intelligent vehicles based on genetic algorithms is proposed. The membership functions and rules base of fuzzy feedback control law are automatically optimized and effectively determined by genetic algorithms. The stability of this fuzzy feedback control system is proved by Popov-Lyapunov. On this basis, the hierarchical control structure of fuzzy feedback control law according to driving speed, and in different layers fuzzy feedback controller is designed via genetic algorithms to optimize the parameters of membership functions and rules, which can effectively improve the adaptive ability lateral control system to the change of speed. The simulation results demonstrate the lateral control system proposed by this paper has better stable and dynamic performance.(4) Study on longitudinal adaptive sliding-mode control of intelligent vehicle using via fuzzy-logic. Firstly, the longitudinal nonlinear dynamic control model of intelligent vehicle is established which is used for the design of longitudinal control system. Secondly, the upper level controller of longitudinal control system is designed by classic PID control method. On the basis of upper level controller, aiming at the large parametric uncertain, external disturbance and time delay properties of longitudinal dynamic control model, the lower level controller of longitudinal control system is designed by adaptive fuzzy sliding mode control method, and the gain coefficients and boundary layers of sliding-mode control are adaptive tuned by fuzzy logic, which can effectively improve the performance and eliminate the chattering phenomenon, consequently, the vehicle speed can be controlled rapidly and exactly by proposed method. Then, the stability of the closed-loop longitudinal control system is proved by Lyapunov method. Finally, the switching criterion between the throttle actuator and brake actuator is presented, which can guarantee the smooth switching between throttle actuator and brake actuator. The simulation results demonstrate that the proposed control algorithm can guarantee the smooth transitions between throttle actuator and brake actuator and implement the rational and accurate control of longitudinal motion in presence of large model uncertainties and external distribution.(5) Experimental study on lateral and longitudinal control algorithms proposed by this paper.To further verify the effectiveness and accuracy of lateral and longitudinal control algorithms proposed this paper, aiming at the typical working conditions of lateral path tracking control and longitudinal speed tracking, the experimental tests of lateral motion control system and longitudinal motion control system of intelligent vehicle are carried out using the prototype vehicle DUTIV-I, respectively. The experimental results indicate that the lateral motion control system and longitudinal motion control system based on proposed method exhibit the better characteristics dynamic of dynamic response, tracking precision and robustness. |
PDF Full Text Request