Control Strategy For Autonomous Vehicles On Off-ramp Route Optimization

Along with great social progress and rapid economic development, technologies in many scientific areas have been evolving by leaps and bounds. Transportation industry is one of the typical areas. At the present stage, in order to realize the goal of Intelligent Transportation System, transportation development is gradually stepping from the informatization into intellectualization, approaching towards establishing a smarter city. In recent years, a series of advanced ideas and technologies to solve social issues, including traffic congestion, road safety, energy consumption and environmental pollution. Among all of these technologies, autonomous vehicle is most representative, which develops fast and successfully. In terms of study on autonomous vehicle, companies and academics pay much more attention to vehicle manufacturing technology rather real scenario simulations, especially lack of research on control strategy for autonomous vehicles on off-ramp route optimization. Therefore, supported by the Chinese National Science Foundation Project Research on microscopic modeling and simulation of freeway traffic flow mixed with autonomous vehicles, this paper proposes a mesoscopic traffic flow model for autonomous vehicle on the basis of off-ramp behavior researches on manual vehicles and field experimental studies on autonomous vehicles. Meanwhile, the optimal strategy for autonomous vehicles with off-ramp route control is simulated by computer, which provides promising approaches for traffic control and management of autonomous vehicles in highway. Here are details:First of all, after introducing some investigations on autonomous vehicles’ development progress and its field trials, this paper comprehensively reviews recent developments in adaptive cruise control modeling and conventional lane-change modeling. Then, the off-ramp traffic flow characteristics are analyzed in both macro and micro levels. The detecting radius is also determined according to the environmental awareness system of autonomous vehicles.Secondly, an original lane-change model based on cellular automata is established for autonomous vehicles. This paper selects Jiang’s adaptive cruise control model as car-following model and divides lane-change models into two types:discretionary lane-change model and mandatory lane change model. In terms of discretionary lane-changing model, a risk factor is introduced in this model. Different from discretionary lane-changing process, the mandatory lane change maneuver consists of five steps:data collection, safe gap computing, measured gap ranking, measured gap classification, lane changing selection and implement.Additionally, following the guidance of hierarchical control of piecewise, control strategy for autonomous vehicles on off-ramp route optimization is described as grading strategy, overlap strategy and mixed strategy. In addition, after establishing a flexible simulation scenario, this paper presents evaluation indexes and a general cost evaluation model, in which parameters are calibrated by the methods of expert scoring and Analytic Hierarchy Process.Finally, multiple sets of simulations on control strategies of off-ramp route optimization are carried out using Matlab tools. In the meantime, this paper evaluates fundamental diagrams for freeway traffic flow, average travel time and lane-changing indexes. In the end, optimal control strategy for autonomous vehicles on off-ramp route is determined after comparing with other strategies in general cost.