Research On State-driven Control And Reservation-oriented Scheduling Mechanisms For Driverless Vehicles

With the advance of multidimensional sensing, intelligent driving, and mobile communication and so on, driverless vehicles have been coming true and playing an important role in the novel Intelligent Traffic System(ITS). Nowadays, with the maturing of autonomous driving technology, the initiative V2I/V2 V collaboration based on vehicular communication technology has been becoming a research hotspot in ITS. In this paper, we mainly focus on the research on the cooperation of heterogeneous vehicles with different priorites, which is prominent differential from current ITS, with the support from NSFC and FFCSA 2012.After analyzing the problem and especially spatial-temporal domain properties of driverless vehicles' motion, we establish formal models for all traffic objects and vehicular kinematics firstly. Then, we propose state-driven and reservation-oriented control mechanisms to represent possible vehicular behaviors during passing-through intersections, designing V2 I communication messages correspondingly. Via combining the properties of vehicular motions and physical environments, we design a centralized reservation-oriented scheduling mechanism. Further, we propose two different priority-inheritence-based scheduling algorithms for single critical section and continuous critical sections, by considering to decrease the priority-inversion phenomena in traffic flows. Finally, we develop a traffic simulation software, in which all designed models, mechanisms, and algorithms in this paper are implemented, and which has played an important role for us to verify and evaluate our study.The results of experiments show that, proposed models and mechanisms are effective to characterize complex traffic situations, and to support the research on collaboration among heterogeneous vehicles. Compared with traditional algorithms, the scheduling algorithms proposed in this paper can effectively decrease delay time of high-priority vehicles, which is the main target of our research. In general, our research work is of practical and referential importance to ITS and the revelant research on multi-robots.