Dynamic Modeling And Distributed Control Of Vehicular Platoon Under The Four-component Framework

Vehicular platoon has potential to singnificantly mitigate traffic congestion, enhance road safety and improve fuel economy. Many existing research only focus on designning tools to analyze a specific vehicular platoon with single communication topology, which lacks a general framework to handle platoons with different information flow topologies. To address aforementioned issue, this thesis proposes a four-component framework for vehicular platoon, and within such framework the closed-loop stability and robust performance are discussed for linear homogenesous platoons. Besides, this thesis further introduces a design process of distributed model predictive control(DMPC) to deal with nonlinear heterogeneous platoons. The results in this thesis can provide tools and lay foundation to address performance analysis and controller design for vechculiar platoons with different information flow topologies.Firstly, a four-component framework is proposed for vehicular paltoons from the perspective of cooperative control of multi-agent system, and the unified models of each component are established under this framework. Such perspective naturally decomposes a platooninto four interrelated components, namely, 1) node dynamics, 2) information flow topology, 3) formation geometry, and 4)distributed controller. This decomposition can provide a unified framework to analyze and quantify the influence of information flow topology, vehicle dyanmics and controller parameters on the platoon performance.Secondly, the closed-loop stability is discussed for linear homogeneous platoons within the proposed four-component framework. Using matrix factorization technique, the platoon dynamics can be decompled, andthus closed-loop stability of platoon is proved to be equivalent to that of several subsystems. The major results are as follows: 1) the stabilizing region of linear controller gains are explicitly established for different information flow topologies by using Routh–Hurwitz stability criteria; 2) through employing matrix eigenvalue analysis, the stability margin of platoons with bidirectional topology is proved to decay to zero as 1 2 when platoon size increases; 3) by employing tools, such asRayleigh-Ritz theorem, this thesis shows how to improve the stability margin of a homogeneous platoon in terms of topology selection and control adjustment from a unified viewpoint;4) adistributed controller synthesis approach is proposed by convertingthe platoon control design into solutions toparametric algebraic Riccati equations.Thirdly, the robust performance is addressed for linear homogeneous platoons under external disturbance with finite energy by analyzing the infinite norm of corresponding transfer functions. The main results are as follows: 1) the robust performance index grows exponentially with increasing platoon size for predecessor-following topology, which is independent with the controller gains; 2) the robust performance index will at least be indepnent with platoon size, if every following vehicle can obtain the leader’s information, resulting the predesessor-leader following topology; 3) the robust performance index at least grows as a polynomial function( 2) of platoon size for bidirectional topology.Finally, a design process of distributed model predictive control(DMPC) is introuduced to deal with the nonlinear heterogenous platoons. The proposed DMPC utilizesthe predicted trajectories of neighboring vehicles to construct the sub-optimization problems, which is well-suited for various information flow topologies. By using the cost function as a Lyapunov candicate, a sufficient condition is proposed to gurantee the asympotatical stability for vehicular platoons with unidirectional information flow topology.