Lane-changing model for autonomous vehicles based on internet of vehicles

Currently, autonomous vehicles are widely favored by the public, but the issue of safe lane-changing remains a major concern. Intelligent lane-changing models face challenges such as the need for large sample data and limitations on overtaking road segments. Therefore, this research optimizes the Internet of Vehicles and constructs a grid-based High-Definition Map lane matching model. Principal component analysis is applied to reduce the dimensions of the lane model, and various road condition information of the current lane is analyzed and matched with cloud data from the Internet of Vehicles. An adaptive lane-changing control system is also included. The optimized Internet of Vehicles model is compared and validated using a public dataset. The model's match with the planned route of the grid-based High-Definition Map exceeds 90% under four different road conditions. In response to winding mountain roads, the model maintains a stable speed, with the standard deviation of the average speed only 0.38m/s. In subsequent simulation experiments, the model safely completes the lane change within about 100 meters. The results show that the proposed lane-changing model can handle complex and high-speed traffic scenarios, safely and stably fulfill the lane change requirements of Autonomous Vehicles, while improving lane change efficiency and reducing energy consumption. This model provides a new approach for the intelligent vehicle industry and ensures road safety.