Development Of Decision-Making For Cooperative Automated Vehicle-Group

Connected and Automated Vehicle(CAV)technology is the trend of transportation and is now in rapid development.Past studies have confirmed that CAV could improve driving safety,comfortability,mobility,and reduce energy consumption.A new type of CAV and Human-driven Vehicle(HV)mixed traffic is upcoming in the future.The CAV would show its great potential in traffic management.Therefore,CAV based multivehicle cooperation technology is now a hot research topic.However,most studies only focus on simple cooperation methods,such as longitudinal platooning.These methods cannot make full use of CAV’s advantages on cooperative driving.Therefore,this thesis proposes a meaningful decision-making method for Cooperative Automated VehicleGroup(CAVG).CAVG is a dense group formed by multi-CAVs on multi-lanes.Compared to the special case of a CAVG in one-dimensional space – a CAV platoon,the CAVG could cooperate CAVs from two dimensions.This not only improves running mobility but also provides transportation management with a new actuation method.A CAVG on multi-lanes could significantly influence traffic flow.Therefore,a CAVG could help with traffic guidance and control.In order to realize the decision-making of CAVG(including vehicle behavior decision-making and motion planning),this thesis makes the following studies.This thesis proposes a decision-making framework for CAVG.This framework considers three different cooperation levels(group,platoon,and vehicle)and three decision-making layers(behavior decision-making,trajectory planning,and control).The framework designs I/O interfaces.Vehicle-based and platoon-based CAVG decision-making logics are designed.Therefore,it could accommodate conventional studies on CAV and platoon.Moreover,this framework consists of a Model Predictive Control(MPC)based motion planning structure which is the core module of decisionmaking.The proposed CAVG decision-making framework may be a good roadmap for future studies.In the CAVG decision-making framework,the decision-making of CAV and platoon is the actuation module of CAVG.Therefore,this thesis makes efforts on the decision-making of CAV and platoon for a better fit to CAVG.A chance-constrained MPC-based motion planning model is proposed for the efficient highway-piloting of Group CAVs.This model could consider the state stochasticity of HV.Field tests and simulation experiments demonstrate that this motion planning model could plan motions with lane-change decisions.In straight-cruising,lane-change,turning at intersections,and curve-cruising scenarios with 60 km/h speed limit,tracking error is less than 12.5 cm and speed error is less than 3.32 km/h.The proposed single-vehicle motion planning method improves control accuracy by 27% compared to a state-of-theart study.The average computation time is 5.15 milliseconds.This indicates that the motion planner is ready for field application.This CAV motion planner provides the decision-making of the CAV-based CAVG with an actuation module and facilitates the field application of CAV technology.To address the need for Platoon Lane-Change(PLC)on highways,this thesis proposed a space domain-based Simultaneous Platoon Lane Change(Si PLC)method and a backward-looking based Cut-in Platoon Lane Change(Ci PLC)method.These two methods could maneuver a platoon to change lane by utilizing a small gap.Vehiclebased simulation and simulation with traffic demonstrate that the proposed two PLC controllers could significantly improve PLC capability and efficiency in low-density and high-density traffic respectively.Moreover,in the PLC maneuvers,string stability is always guaranteed.The Si PLC and Ci PLC methods could address the PLC needs and benefit its field applications.They also provide the decision-making of the platoonbased CAVG with actuation modules.Supported by the CAV motion planner in CAV and HV mixed traffic,the Si PLC controller,and the Ci PLC controller,a moving-cell-based CAVG decision-making model is proposed in this thesis.It is designed for the efficient cruising of CAVG on highways.CAV motion planning and cooperative lane-change control are the actuation module of the proposed CAVG decision-making model.This model could optimize behavior for CAVs in groups and maneuver the CAVG overtaking slow-moving HVs.The proposed CAVG decision-making model realizes the decision-making process of CAVG,including behavior decision-making,planning,and control.It is pioneering research in CAVG decision-making and gives chance to CAVG-based traffic management.