A Hierarchical Approach to Trajectory Planning for Autonomous Vehicles on Curvy Roads

Zhao, Fuzhou; Huang, Heye; Wang, Xuemei; Han, Ling; Ye, Fei; Cui, Mingyang; Zhong, Shan · 2025 · Research Square

DOI: 10.21203/rs.3.rs-6270718/v1

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the dual challenges of decision-making efficiency and trajectory maneuverability for autonomous vehicles navigating curvy roads in dynamic traffic environments. While existing trajectory planning methods often prioritize collision safety, they frequently neglect the computational burden of high-speed planning and the kinematic constraints required for passenger comfort and vehicle stability. To resolve these issues, the authors propose a three-layer hierarchical trajectory planner that integrates an improved dynamic programming (DP) algorithm with adaptive cruise control (ACC) strategies to generate smooth, kinematically compliant paths. The methodology employs a Frenet frame coordinate system to model traffic scenarios, decoupling longitudinal and lateral motion. The first layer, the coarse trajectory layer, utilizes a novel "jump node DP algorithm." This algorithm optimizes the traditional DP approach by identifying and skipping grid nodes that overlap with the contours of surrounding vehicles, thereby eliminating redundant collision checks and cost calculations. The second layer, the reference trajectory layer, applies joint interpolation to the coarse trajectory to ensure continuity in heading angle and acceleration. The third layer, the alternate trajectory layer, iteratively replans trajectory segments with high curvature using an ACC strategy to follow leading vehicles, deferring overtaking maneuvers until the vehicle traverses specific waypoints, after which the improved DP strategy resumes. Simulation experiments demonstrate significant improvements in both efficiency and maneuverability. The jump node DP algorithm achieved an average speedup of 30.6% compared to traditional DP algorithms, reducing CPU and memory overhead by omitting calculations for overlapping nodes. Furthermore, the hierarchical approach enhanced the practicality of the planned trajectories; the minimum radius of curvature in the planned paths increased by up to 12.7%, indicating smoother and safer maneuvering that better aligns with vehicle kinematic constraints. Dynamic simulations in complex traffic environments confirmed the effectiveness of the proposed framework, which also provides redundant failure protection by allowing lower-layer trajectories to serve as fallbacks. The significance of this work lies in its balanced approach to trajectory planning, addressing the often-overlooked trade-off between computational speed and vehicle dynamics. By integrating efficient search algorithms with kinematic smoothing and adaptive control strategies, the proposed method offers a robust solution for autonomous driving on complex road geometries. The findings suggest that hierarchical planning architectures can substantially improve the real-time performance and safety of intelligent vehicles, paving the way for more reliable autonomous navigation in intricate traffic scenarios such as intersections and merges.

Key finding

The proposed hierarchical trajectory planner improves decision-making efficiency by up to 30.6% and increases the minimum curvature radius by up to 12.7% compared to conventional methods in simulated environments.

Methodology

simulation_modeling

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-28.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• road geometry