A dual-dimensionality reduction strategy for optimization-based parallel parking path planner

Optimization-based parking path planner has been attracting much more attention due to its capability to generate the optimal path rather than a feasible one. But when solving this optimal problem, it is difficult to make a good balance between the computational accuracy and efficiency. Because dens...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Expert systems with applications Jg. 263; S. 125781
Hauptverfasser: Hu, Qiuxia, Ma, Jie, Zhan, Guanglun, Gao, Feng
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier Ltd 05.03.2025
Schlagworte:
Automated parking
Nonlinear programming problem
Obstacle avoidance
Optimal control
Path planning
Nonlinear programming problem
Automated parking
Path planning
Obstacle avoidance
Optimal control
ISSN:0957-4174
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Optimization-based parking path planner has been attracting much more attention due to its capability to generate the optimal path rather than a feasible one. But when solving this optimal problem, it is difficult to make a good balance between the computational accuracy and efficiency. Because dense discretization points are necessary for accuracy, which conversely leads to a high dimensionality of optimization space. To deal with this challenge, this paper proposes a dual-dimensionality reduction strategy to accelerate the numerical optimization process of parallel parking path. One is to reduce the number of optimized parameters by approximating the optimal path with multiple circular arcs connected end to end, according to the numerical statistical analysis of the geometric features of the optimal path. The other is to reduce the number of nonlinear and non-convex constraints for obstacle avoidance by analyzing the geometry relationship between the parameterized arcs and the boundary lines of obstacles. This point-over-arc obstacle avoidance constraint can ensure that the planned path is collision-free by only checking the limitations at the endpoints of the circular arcs. The effectiveness of the proposed method on the improvement in computational efficiency is validated by several comparative simulation and experimental tests. Compared with the traditional numerical optimization methods, the presented algorithm improves the computational efficiency by nearly 40 times on average.
ISSN:0957-4174
DOI:10.1016/j.eswa.2024.125781