Purchase this article with an account.
Brett R. Fajen, Nicholas O. Beem, William H. Warren; Route selection emerges from the dynamics of steering and obstacle avoidance. Journal of Vision 2002;2(7):418. doi: 10.1167/2.7.418.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Previous experiments on elementary steering and obstacle avoidance behavior in humans were used to develop a dynamical model of on-line route selection in simple and complex scenes (Warren, Fajen, & Belcher, VSS 01). The model describes how turning is influenced by the current distance and heading angle of goals and obstacles. In the present experiments, we tested several model predictions about route selection by having participants walk to a goal through specific configurations of obstacles in a 12 × 12 m virtual environment. They wore a head mounted display (60 × 40 deg, 50 ms latency) and head position and orientation were recorded by a sonic/inertial motion tracking system. In Experiment 1, a pair of obstacles was presented, one on either side of the initial heading to the goal and at different distances. When the initial angle between the far obstacle and the goal was small, participants favored an outside route around the far obstacle. As this angle was increased, participants switched to an outside route around the near obstacle, and then to a route between the two obstacles. The same sequence of bifurcations was predicted by the model. In Experiment 2, we created a “local minimum” by presenting a cul-de-sac, an array of obstacles arranged in an arc. Participants were instructed to walk to the goal by either passing through the array or detouring around it. The arc length and the width of the gap between obstacles were varied independently. Participants were more likely to walk directly through the cul-de-sac as arc length increased (creating a larger barrier) and as gap width increased. These data were consistent with the model, which similarly avoided local minima. The results demonstrate that human route selection in complex scenes can be understood in terms of the competition between on-line steering and obstacle avoidance strategies.
NIH R01 EY10923, NIH KO2 MH01353, NSF DGE 9870676
This PDF is available to Subscribers Only