Purchase this article with an account.
Justin Owens, William Warren; Can people learn to anticipate obstacle motion when necessary to avoid collision?. Journal of Vision 2008;8(6):1156. doi: 10.1167/8.6.1156.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Avoiding moving obstacles is critical to our ability to safely guide locomotion through a complex environment. In many circumstances (e.g. rapid movement), it is advantageous to anticipate which environmental objects are likely become obstacles. Fajen & Warren (2003) proposed a dynamical model of on-line steering control that requires no higher-level path planning or anticipation. Although participants can learn to anticipate a target's motion with repeated exposure to one trajectory, this is not the case for two trajectories, which yield behavior similar to the model's predictions (Owens & Warren, VSS 2006). Participants also fail to anticipate the trajectories of one or two moving obstacles, even when they are cued by color and shape (Owens & Warren, VSS 2007).
The current study investigates whether additional constraints can induce participants to anticipate obstacle trajectories and preemptively avoid collision. Participants walk in the VENLab, a 12m × 12m virtual environment with a head-mounted display (60 deg H × 50 deg V) and a sonic/inertial tracking system (latency 70 ms). As in the previous study, participants walk to a stationary goal 6m away. A potential obstacle initially approaches on a path parallel to the participant's, at a speed equal to her current walking speed. When the participant crosses a fixed threshold in space, the obstacle veers rapidly onto a constant secondary trajectory. The terminus of this second trajectory is a point immediately beyond the threshold where the participant would be were she walking directly from the starting position to the goal. If the participant does not learn to deviate before the obstacle veers, a collision will occur. In a control condition, the obstacle is initially stationary at the veering point. We predict that the increased demand of collision avoidance will cause participants to learn to anticipate moving obstacles that pose a potential threat.
This PDF is available to Subscribers Only