Intelligent Vehicle Longitudinal Trajectory Planning Based On Cut-in Behavior Recognition

The explosion of the information revolution has brought about rapid development and progress in fields such as computers,artificial intelligence,communication technology,and sensor technology,providing new development ideas for traditional transportation tools.As an interdisciplinary field integrating multiple high-level technologies,autonomous driving technology can provide more solutions for future industrial development.Among them,the decision planning module of autonomous driving vehicles is one of the core components,which should be able to make real-time decisions and plan smooth,safe,reasonable,comfortable,and obstacle-avoiding trajectories.This article derives the conversion relationships between world coordinates,vehicle body coordinates,and Frenet coordinates based on commonly used coordinate systems in autonomous driving systems and analyzes the role of Frenet coordinates in trajectory planning.By decoupling the road environment horizontally and vertically and simplifying the road model,the computational efficiency can be improved.At the same time,the entire process of the trajectory planning algorithm is modeled and deduced based on Frenet coordinates,including sampling the lateral and longitudinal trajectories,generating trajectory clusters through polynomial fitting,establishing a cost function for lateral and longitudinal trajectories by considering evaluation indicators such as reachability,comfort,and safety,and using it to evaluate and filter trajectory clusters,and finally synthesizing the trajectory and transforming it into the world coordinate system.The collision detection of trajectories is conducted using the Separating Axis Theorem and OBB-type bounding boxes,and the trajectory with the minimum cost is selected as the output of the decision planning module and sent to the control module for tracking control.This article explores the important role of Frenet coordinates in autonomous driving trajectory planning,improves computational efficiency by simplifying the road model,and models and deduces the trajectory planning algorithm,ultimately achieving the selection of the trajectory with the minimum cost as the output of the decision planning module.In addition,this article analyzes the vehicle motion parameters for the common cut-in scenario on actual roads and adopts multidimensional Kalman filtering to smooth and filter the unstable obstacle information from the perception module.A lateral safety distance model is established based on lateral velocity and position for the autonomous driving system to identify the cut-in intentions of obstacles.Based on an analysis and modeling of four commonly researched safety distance models,this article selects the Active Fuzzy Safety Measure as the longitudinal safety distance model and combines it with the cut-in time of obstacles to derive the final state of the autonomous driving vehicle’s longitudinal decision-making and trajectory planning.Therefore,this article uses multidimensional Kalman filtering to smooth the obstacle information from the perception module and adopts the Active Fuzzy Safety Measure as the longitudinal safety distance model to derive the final state of the autonomous driving vehicle’s longitudinal decision-making,thus conducting trajectory planning.Finally,a joint simulation platform and real vehicle testing scenario are established.In the simulation testing,different vehicle speeds and cut-in distances are designed to verify the optimized algorithm before and after optimization.After analyzing the results,the initial braking time of the optimized algorithm is earlier than that of the unoptimized algorithm,and the average,maximum deceleration,and average jerk are smaller than those of the unoptimized algorithm.This confirms that the optimized algorithm can plan a forward and smooth braking motion curve in the face of the cut-in scenario of the front vehicle,improving driving safety and ride comfort.This research is of great significance for the trajectory planning of autonomous driving systems,as it can not only enhance driving safety but also improve ride comfort.Moreover,this research provides an effective solution for some cut-in scenarios on actual roads.