| Regards to the increasingly severe traffic problems，it is an important issuedemanding prompt solution that how to accurately cognize the traffic phenomena in reallife with scientific methods, and how to relieve traffic jam, reduce emissions/energyconsumption and improve driving safety with advanced technology development oftransportation system. Currently, the informatization and intelligence of transportationhas become the main development direction of the modernization of internationaltransportation system, which is the front area in subject of contemporary science andtechnology. However, the traditional transportation system has not yet achieved fullcoordination and optimization as a result of the lack of extensive interconnection,intercommunication and interoperability among the process of perception,communication, computation and control. And with the emergence of cyber-physicalsystem (CPS) and traffic flow theory, a new way to solve the above problems from theviewspoint of advanced technology and accurate perception arises. Therefore, it is withimportant theoretical and practical significance to study the traffic cognition method andrelated problems in the transportation cyber physical systems (T-CPS) environment.This dissertation focus on the T-CPS oriented microscopic traffic cognitivetechniques and related key problems. Specially, a service-oriented architecture for theT-CPS is put forward according to the lack of extensive interconnection,intercommunication and interoperability among the process of perception,communication, computation and control of the traditional transportation system. Thenthe systematic discussions of optimal velocity function, stability analysis, nonlinearanalysis and zero-dynamics of microscopic car-following model are conducted based onmicroscopic traffic flow theory. Moreover, the feedback control of traffic congestion isinvestigated from the system and control perspective, and thereby, a called multipleheadway, velocity&acceleration difference (MHVAD) car-following model is proposedtaking into account the comprehensive information in the T-CPS environment. In theend, theoretical analysis and numerical simulation both prove the validity of MHVADmodel.The main works of this dissertation are as follows:①Considering the interaction and feedback between the physical system andcyber system of CPS, the basic architecture of T-CPS including perception, communication, computation, control and service is proposed.Since the traditional transportation system has not yet achieved full coordinationand optimization as a result of the lack of widespread interconnection,intercommunication and interoperability. And CPS is a multi-dimensional complexsystem, integrating of cyber element and physical element through the3C (Computation,Communication, Control) technology. Its main characteristic is achieving the preciseperception and effective control of the physical system through the interaction andfeedback between the cyber system and physical system. So the dissertation introducesthe CPS into transportation system and puts forward a service-oriented architecture forT-CPS, and then explains the function of corresponding layers, finally discusses somekey techniques and applications of T-CPS to provide theory support for next generationof ITS.②Regards to the microscopic cognitive technique for trafficphenomenon——car-following model, the systematic discussions of the key issues ofthe following model, such as optimal velocity function, stability analysis, nonlineardynamic analysis and the zero dynamics are carried out.Car-following model attracted more attention is one of the main approaches tocognize the revolution of traffic state from the microscopic viewpoint in T-CPS. Andcurrently, many car-following models are improved as a function of OV model,nevertheless, the key of OV model lies on the optimal velocity function. Therefore, it isof importance to study the cognitive technique based on car-following model with thecharacteristics of optimal velocity function.The stability analysis of car-following model is one important aspect in the trafficcognitive based on car-following model. And the local stability and the asymptoticstability of car-following model are systematic discussed in this dissemination. Inconsideration of the strong coupling and nonlinear characteristics of traffic system, toanalyze the stability effectively, a lemma of Lyapunov function construction for oneclass of second-order system is put forward and proved according to the related controltheory, and then, the Lyapunov stability of car-following model is discussed.To reveal the nonlinear characteristics of traffic system, on one hand, the mKdVequation of car-following model is derived through the small variation methodaccording to the traditional approach; while on the other hand, a new attempt to obtainthe zero dynamics equation of car-following model is carried out with emphasis on theconcept of zero dynamics. ③The feedback control of traffic jam based on car-following model isinvestigated from the system and control viewpoint.The traffic system is essentially a complex system. According to the fundamentallaws of physics, force is the reason for the movement of objects, while the accelerationis the reason for changing the state of the movement of objects. To conduct a study fromthe system and control perspective, the acceleration information is designed as thefeedback signal, and the feedback control of traffic jam based on car-following model isdiscussed based on the FVD model. And smulation results show the effectiveness of theproposed method in the final.④A MHVAD car-following model is proposed in consideration of thecomprehensive information in T-CPS environment.To cognize the traffic phenomenon accurately, a MHVAD car-following model isput forward taking into account the comprehensive impacts of the number, position,velocity difference and acceleration difference of preceding vehicles. And the stabilityof MHVAD model is conducted in detail. Then the dynamic performance of MHVADmodel is discussed from the start process, stop process and revolution process.Compared with the existing models, the simulation results show that the performance ofMHVAD model is better than that of those models, which can provide the basis fortraffic state cognitive accurately.In conclusion, the basic architecture of T-CPS is proposed, and regards to themicroscopic cognitive technique for traffic phenomenon——car-following model, thesystematic discussions of its optimal velocity function, stability analysis, nonlinearanalysis are conducted, moreover, the zero dynamics of car-following model is made,then, the feedback control of traffic jam is studied, and in the end, a MHVADcar-following model is proposed and theoretical analysises and simulation resultsvalidate the effectiveness of the work. |
PDF Full Text Request