Scholarlink

<p>Instantaneous trajectory planning is crucial for the autonomous navigation of mobile robots. Among existing methods, the Dynamic Window Approach (DWA) serves as a widely adopted benchmark. However, when applied to omnidirectional robots with high aspect ratios, conventional DWA suffers from limitations such as kinematic mismatches, overly conservative geometric approximations, and non-adaptive decision-making. To tackle these challenges, this paper first extends the conventional 2D velocity space to generalized planar velocity space to fully utilize the platform’s kinematics. Second, a precise rectangular footprint model is developed to replace the overly conservative circular envelope. Finally, a lightweight, real-time risk-based mechanism is developed for adaptive weight tuning. Extensive comparative simulations in various typical scenarios validate the significant advantages of the proposed improved DWA. The results show that in scenarios with narrow passages, the precise geometric model shortens the path length by 49.0% while avoiding collisions. The real-time experiment on a testbed across sparse and dense environments is performed too, yielding near-optimal traversal efficiency by the adaptive weight tuning mechanism.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Instantaneous trajectory planning of omnidirectional mobile robot in generalized planar velocity space by dynamic window approach

  • 期刊论文

  • 发表日期：

    2026年5月30日

查看全文
  • Zheng Xiao,
  • Yafeng Tian,
  • Silu Chen,
  • Tianjiang Zheng,
  • Chi Zhang,
  • Guilin Yang

摘要

Instantaneous trajectory planning is crucial for the autonomous navigation of mobile robots. Among existing methods, the Dynamic Window Approach (DWA) serves as a widely adopted benchmark. However, when applied to omnidirectional robots with high aspect ratios, conventional DWA suffers from limitations such as kinematic mismatches, overly conservative geometric approximations, and non-adaptive decision-making. To tackle these challenges, this paper first extends the conventional 2D velocity space to generalized planar velocity space to fully utilize the platform’s kinematics. Second, a precise rectangular footprint model is developed to replace the overly conservative circular envelope. Finally, a lightweight, real-time risk-based mechanism is developed for adaptive weight tuning. Extensive comparative simulations in various typical scenarios validate the significant advantages of the proposed improved DWA. The results show that in scenarios with narrow passages, the precise geometric model shortens the path length by 49.0% while avoiding collisions. The real-time experiment on a testbed across sparse and dense environments is performed too, yielding near-optimal traversal efficiency by the adaptive weight tuning mechanism.