Studies On Integrated Vehicle Controls Based On Constrained Optimization

Vehicle electronic control not only represents the megatrends of automotive technologies in the future, but also the key responses to the challenges faced by the automobile industry in energy saving, environmental protection, safety and congestion. As electronic control systems become increasingly complex, they often possess strong dynamic coupling characteristics with multiple control objectives that often are contradicting between each other. In addition, these systems are often over actuated, dynamically constrained with strong nonlinear characteristics. They share information resources, interact with tire and human, and tend to exhibit a characteristic of strong coupling and conflict. Therefore, the integration of vehicle electronic control systems becomes increasingly important in order to improve the control efficiency, and to the maximum extent, to enhance the overall vehicle performance.Integrated vehicle control has been a focused issue in academy and industrial for a long time. Although the studies of integrated vehicle control have fruitful progress in recent years, many different research methods and technology roadmaps are endlessly proposed expresses the immaturity of the technology in this area, especially in systematized design theory and method.Traditional vehicle electronic control is still mainly based on feedback control, which leads to each subsystem independent, thus results in difficulty to be a whole system. That presents more challenges for integrated control. With the development of the vehicle electrification, and emerging of many kinds of electronic control methods and X-by-wire actuators, the over actuated characteristic of the control system becomes widespread, which provide a new subject for integrated control. Traditionally, allocation-based integrated control methods are difficult to not only adapt to the complicated dynamic driving process, but also represent the dynamic changes and couple process among the actuators, tires and roads. The constrained optimization based control allocation methods have been aroused general concern in recent years, and those methods have been applied in vehicle control area. However, the consideration of constrains is still simple in many researches, and the constrains are often limited in the adhesion constrains between tires and roads. Although the actuator constrains are considered in some researches, they are simplified as bounds of the tire forces in general. The coupling properties between longitudinal and lateral tire forces are also rarely considered. Therefore, the solving of optimization may not be achieved, and the actuators cannot be fully utilized. Besides, the stability and efficiency of the numerical algorithms are very crucial to ensure the real time requirements, which is scarcely mentioned in the existing researches.The actuators in vehicle electronic control system contains drive system, brake system and steering system, which are inter-coupling and mutually restrictive. The dynamic response characteristics of the three systems are quite different. Therefore, the dynamic characteristics are critical factors in design of control allocation method for over actuated system. However, the allocation methods proposed before are mostly based on the steady-state characteristics of the actuators, and neglect the transient characteristics of actuators and the transient response differences among them, which results in the dynamic allocation error of the control system, and have a negative effect on the performance and efficiency of the control.As more and more electronic systems and X-by-wire systems are applied in the vehicle, the reliability of vehicle control system becomes more and more important, which ceaselessly promotes the study on the fault-tolerant control methods. Although the fault-tolerant control method based on the over-actuated redundant control allocation has been widely concerned, the failure modes of the actuator, the effects of failure on the performance of control system are not well considered in present researches, which directly results in certain randomness and blindness in the redundant allocation when designing the fault-tolerant control method.In this paper, an integrated vehicle control method based on multi-objective optimization under complex constraints has been proposed by establishing complex constrained relationships in vehicle electronic control system and describing the multiple objectives in mathematics. Based on that, an over actuated control allocation method has been proposed by considering the dynamic characteristics of each actuator. The failure effects of the actuators have been studied by modeling and analyzing the failures of actuators. A constraints optimization based fault tolerance control method has been proposed by the fault tolerance analytical method.Firstly, an integrated vehicle control method based on multi-objective optimization under complex constraints has been proposed, which includes recognition of driver intention based on the model of vehicle dynamics response, vehicle motion control based on inverse vehicle model and state feedback, control allocation method based on constraints optimization, and actuator control method based on inverse tire model. The influence of the choice of the optimization objective on the result of control allocation has been analyzed. The constraints of tire force, which take into account the coupling characteristics of tire force in longitudinal and lateral, have been derived and simplified from quadratic to linear. Reallocation pseudo inverse algorithm and active set algorithm have been derived to solve the constrained optimization of tire forces, which lays the foundation for the implement of the proposed method. Finally, simulations have been carried out to compare with the conventional stability control methods. The results show that the proposed allocation method not only improves the response speed of the control system, but also has more accurate and efficient performances.Secondly, the actuators’ dynamic characteristics have been analyzed and studied systematically by establishing the dynamics model of drive system, brake system, and steering system. A tire force allocation method has been proposed by allocation in frequency domain and compensation of the dynamic characteristics. Finally, the simulation results are compared with that of the traditional control allocation methods without considering the dynamic characteristics of actuators. The results show that the proposed allocation method improves transient response and reduce the control effort of the system.Third, a Fault Tolerance Control(FTC) method has also been proposed. The effect of actuators’ failure on the control performance has been studied by modeling and analyzing of actuators’ failure. A constraints optimization based fault tolerance control method has been proposed by fault tolerance analytical method. To verify the fault-tolerant control method proposed in this paper, typical simulations have been carried out, including one wheel brake failure, one wheel free steer error, and one wheel fixed steer error. The results show that the failure model can well express common failure of actuator, the fault detect method can determine whether the actuator is failure immediately, and the proposed FTC method can handle the failure effectively and avoid the degradation of the performances.A real vehicle testing platform, including mechanical system, electricity and electronics system, control system and data sampling system, has been established to verify the proposed constrained optimization-based IVC method. This vehicle is equipped with four in-wheel motors and electro-hydraulic brake system. The control algorithm can be verified and iterated rapidly by using Micro Auto Box II as a rapid control prototyping. A general control logic has been developed to verify some special control algorithms. ISO-3888-2:2002 test and FMVSS126 test have been carried out. The results show that IVC method can effectively improve the vehicle stability, and the numerical optimization algorithm can be also solved in real time.The main innovations of this paper are:(1) An integrated vehicle control method based on multi-objective optimization under complex constraints has been proposed in this paper, which contains complicated models of actuators’ constraints, constraints between actuators and tires, and adhesion between tire and road surface. A novel numerical algorithm has been also proposed which can efficiently solve the constrained optimization problem in real time.(2) A tire force allocation method which considers the actuators’ dynamic characteristics has been proposed, and the actuators’ dynamic characteristics have also been analyzed by modeling the subsystems dynamics, which lays the foundation of improving the integrated control performances.(3) A fault tolerant control based on the failure analysis of actuators has been proposed. The effects of actuators’ failure on the control performance have also been analyzed by modeling actuators’ failure, and the fault tolerant control method is also proposed by constrained optimization.