Vehicular Adaptive Cruise Control On Curved Road

Adaptive Cruise Control (ACC) system has been believed to reduce driver’sworkload and improve driving safety effectively. However, existing ACC targetdetermination methods by radar cannot detect preceding vehicles properly on a curvedroad. And when ACC vehicle is running at a high speed to keep car-following on acurved road, it will even lose its lateral stability. A Direct Yaw-moment Control (DYC)system is always adopted to improve vehicle lateral stability. However, ACC and DYCare reversely interactive with each other in some driving situations. To address thoseissues, this thesis proposes a Curving ACC system that is coordinated with DYC andinvestigates its key technologies.To improve the environment perception of ACC, a new algorithm for radar-basedtarget vehicle identification and tracking on curved roads for ACC applications isstudied firstly. The identification between target curve-entry/exit and lane-change isaehieved based on the different relationship between relative velocity and azimuth angle.And the distinction of the host-lane preceding vehicle from the adjacent-lane one whiledriving through a curve is realized based on the deviation of the measured lateraldistance from a theoretical value.Then, to maximize longitudinal car-following capability while ensuring the lateralstability on curved road, a DYC based Curving ACC coordination controller is designedunder the framework of Receding Horizon Control theory, focusing on the generictechnologies study including predictive model building based on integrated vehiclelongitudinal/lateral dynamics and generalized inter-vehicle dynamics, performanceindex design considering the contradictions between vehicle longitudinal car-followingand lateral stability, and the derivation of practical problems as poor robustness to themodel uncertainties and high on-line computing complexity for receding optimizationalgorithm application on vehicles.Moreover, to execute the control output of abovementioned coordination controlleras desired longitudinal acceleration and desired yaw moment, the technology for sharingthe same actuator structure between ACC and DYC is studied and the servo controllerof Curving ACC is modified so as to reduce or eliminate the unexpected behavior ofshared actuators. To verify the validity of the Curving ACC system，a series of driver-in-the-looptests are carried out under three specified speed profiles of preceding vehicle andadhesion conditions on curved road and the test results confirm the followings. Thetarget vehicle around a curve can be properly identified and tracked by the proposedradar-based curving target vehicle identification and tracking method. The CurvingACC system can not only improve the lateral stability and weaken the impact of DYCextra brake on ACC longitudinal car-following performance, but also enhance thecar-following capability in limit conditions on curved road.