Work

A Semantic-Enhanced Hierarchical Trajectory Planning Framework with Spatiotemporal Potential Field for Autonomous Electric Vehicles

Year: 2025

Type: article

Abstract: Trajectory planning for intelligent connected vehicles (ICVs) must simultaneously address safety, efficiency, and environmental impact to align with sustainable development goals. This paper proposes a novel hierarchical trajectory planning framework, designed for intelligent connected vehicles (ICVs) that integrates a semantic corridor with a spatiotemporal potential field. First, a spatiotemporal safety corridor, enhanced with semantic labels (e.g., low-carbon zones and recommended speeds), delineates the feasible driving region. Subsequently, a multi-objective sampling optimization method generates candidate trajectories that balance safety, comfort and energy consumption. The optimal candidate is refined using a spatiotemporal potential field, which dynamically integrates obstacle predictions and sustainability incentives to achieve smooth and eco-friendly navigation. Comprehensive simulations in typical urban scenarios demonstrate that the proposed method reduces energy consumption by up to 8.43% while maintaining safety and a high level of comfort, compared with benchmark methods. Furthermore, the method’s practical efficacy is validated using real-world vehicle data, showing that the planned trajectories closely align with naturalistic driving behavior and demonstrate safe, smooth, and intelligent behaviors in complex lane-changing scenarios. The validation using 113 real-world truck lane-changing cases demonstrates high consistency with naturalistic driving behavior. These results highlight the framework’s potential to advance sustainable intelligent transportation systems by harmonizing safety, comfort, efficiency, and environmental objectives.

Source: World Electric Vehicle Journal

DOI: 10.3390/wevj17010022

Authors: Yang Zhao, Du Chigan, Qiang Shi, Yingjie Deng, Jianbei Liu

Institutions: Chang'an University, CCCC Highway Consultants (China)

Language: English

Topic: Robotic Path Planning Algorithms

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

SDGs: Industry, innovation and infrastructure, Sustainable cities and communities, Affordable and clean energy

Open Access status: gold

Estimated APC paid: $1,689