Scholarlink

<p>In rough terrain, walkers must use vision to select viable paths and to find suitable footholds within the time frame allowed by the gait cycle. How is this visuo-motor loop controlled? Recent work demonstrated that walkers look closer to the body in complex terrain, presumably to improve visual information about the terrain structure. However, because the 3D geometry of the terrain was not known, precision of the estimates of gaze and foothold location was limited. Using a 3D modeling technique called photogrammetry, we were able to create 3D representations of the terrain using a head-mounted scene camera. This allows more accurate localization of both gaze and footholds. The goal of the current paper was to measure how closely gaze fell relative to footholds. On medium and rough terrains, gaze was centered around individual steps, with a standard deviation of about 0.25&#xa0;m, indicating the primary role of gaze was visual information for foothold selection, as well as for longer-term path planning. The spread of the distributions suggests that information from the parafoveal retina is often adequate for guiding foothold selection, even in highly irregular terrain.</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

Visual control of walking using terrain reconstructions

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月19日

查看全文
  • Daniel P. Panfili,
  • Karl Muller,
  • Kathryn Bonnen,
  • Mary M. Hayhoe

摘要

In rough terrain, walkers must use vision to select viable paths and to find suitable footholds within the time frame allowed by the gait cycle. How is this visuo-motor loop controlled? Recent work demonstrated that walkers look closer to the body in complex terrain, presumably to improve visual information about the terrain structure. However, because the 3D geometry of the terrain was not known, precision of the estimates of gaze and foothold location was limited. Using a 3D modeling technique called photogrammetry, we were able to create 3D representations of the terrain using a head-mounted scene camera. This allows more accurate localization of both gaze and footholds. The goal of the current paper was to measure how closely gaze fell relative to footholds. On medium and rough terrains, gaze was centered around individual steps, with a standard deviation of about 0.25 m, indicating the primary role of gaze was visual information for foothold selection, as well as for longer-term path planning. The spread of the distributions suggests that information from the parafoveal retina is often adequate for guiding foothold selection, even in highly irregular terrain.