Study On The Key Technology Of Vehicle Longitudinal Following Stability And Active Safety Control

With the increase of automobile inventory, the problems of trafc accidents andtrafc congestion become more and more serious. It has very important signifcance toconduct research on assisting the driver with automatic drive control, which will helpsolve the current trafc problems by increasing vehicle active safety and road efciency.The vehicle longitudinal following system is one of the most important subsystemsin automated highway systems. Meanwhile, it is also the basis for advanced active safetysystems such as intelligent adaptive cruise control, autonomous driving, vehicle platooncontrol, etc. It can enhance the safety and efciency of highway systems by groupingvehicles in a platoon with automatic control. However, there are still some problems in theresearch which need to be solved. Firstly, because the number of vehicles entering the roadcould not be determined, an infnite dimensional complex interconnected model is neededto describe this system. As the stability of a single vehicle and the string stability of aplatoon can not guarantee the collision avoidance of the whole system, we must introduceextra constraint to the states, which increases the difculty of control design. Secondly,as the vehicle is subject to strong nonlinearity and external unknown disturbance efect,modeling uncertainties can not be avoided. The traditional linearization method haslimitations, the nonlinear and uncertainty in the system must be considered. This alsochallenges the the control design. Therefore, improper handling of the above two classesof problems may cause the instability of the whole system, which leads to serious trafccongestion and trafc accidents.For the problems above, this paper conducts research on the key technology of thestability and active safety control for vehicle longitudinal following system. The maincontents and novelties are as follows:(1) This paper proposes a global active collision avoidance decentralized robust con-trol design method for vehicle longitudinal following systems. This method frst givesa class of injective function which transforms the whole collision avoidance space errorstate space to a new state space which contains all real numbers. Then, based on theLyapunov-Minimax approach, a decentralized robust controller is designed which rendersthe transformed system uniform bounded and uniform ultimate bounded (i.e., practicalstability) in the new space. The active collision avoidance and string stability of theconventional system can be deduced. Compared to the traditional method which needsthe initial condition of the system close to the equilibrium point, this method makes the resulting controlled system collision avoidance guaranteed under any safe initial condition.(2) This paper proposes to use the fuzzy set theory to describe the uncertaintieswith a optimal design method for the robust control gain in the fuzzy sense for thevehicle longitudinal following system. A new class of deterministic decentralized robustcontroller is suggested to suppress the fuzzy uncertainty in the system. The controlis proven to render each transformed error dynamics globally practically stable. Thisdeterministic performance guarantees, in the worst case, certain characteristics, whilethe fuzzy performance incorporates an optimal consideration. An optimal design problemassociated with the control is formulated, which enables us to choose the optimal gainfor the robust control. It is proven that the global solution to the optimal problem existsand is unique.(3) This paper proposes a decentralized robust control design method to tackle twoseemingly contradicting objectives for the vehicle longitudinal following system, namelycollision avoidance and compactness. First, a creative bijective transformation is made onthe error dynamics of each vehicle in the platoon. Then, a decentralized robust controlleris designed which renders the transformed system practically stable. Compared withthe traditional control method which can not quantify the derived results, this methodmakes the space error in a given strict interval under any safe initial conditions. Thishelps increase road efciency by decreasing the space error of each vehicle in platoons.(4) This paper proposes a decentralized adaptive robust control design method forvehicle longitudinal following system with a optimal design method for the adaptive ro-bust control in the fuzzy sense. The fuzzy set theory is used to express the uncertaintyin the system. A new class of deterministic adaptive decentralized robust controller. Thecontrolled system is proved to have the properties of two parts. The deterministic per-formance including uniform boundedness and uniform ultimate boundedness guaranteesactive collision avoidance in the worst case. The fuzzy performance assures the accumu-lative system performance and control efort are optimized in the fuzzy sense. The globalsolution to the optimal problem is proven to be exist and unique.