Scenario Digitalization Autonomous Driving Based on Digital Twin Maps

Constrained by the complexity of mixed traffic environments, autonomous driving is currently operational only within certain enclosed scenarios and is not yet equipped to cater to the daily commuting needs of the populace. In response to this challenge, we propose a novel approach to autonomous driving, referred to as scenario digitalization autonomous driving. Scenario digitalization autonomous driving commences with the utilization of high definition map to render a static digital repre-sentation of the driving scene. This is then coupled with vehicle-to-cloud communication to construct a real-time digital twin scene of both the vehicle and its environment in the cloud. The allocation of road rights for autonomous vehicles is subsequently managed through a cloud control dispatch platform. This process creates a structured, efficient digital driving environment that fundamentally circumvents the endless predicament of game-theoretical scenarios potentially encountered by autonomous vehicles, thereby ensuring the smooth operation of the traffic network. In conclusion, the feasibility of scenario digitalization autonomous driving is corroborated through two experiments. These tests confirm the efficacy of the cloud control dispatch platform in terms of its robust scheduling capabilities and its precise interventions for safety, thereby endorsing the viability of this novel approach to autonomous driving,