Related Experiment Videos
Behavioral dynamics of steering, obstacle avoidance, and route selection.
Brett R Fajen1, William H Warren
1Department of Cognitive and Linguistic Sciences, Brown University, USA. fajenb@rpi.edu
Summary
Human navigation in virtual environments is predictable. Steering dynamics, influenced by goals and obstacles, determine route selection, suggesting online adjustments rather than explicit path planning.
Area of Science:
- Human-computer interaction
- Robotics
- Cognitive psychology
Background:
- Understanding human navigation is crucial for designing intuitive virtual environments and autonomous systems.
- Predicting movement in complex environments requires modeling steering and obstacle avoidance behaviors.
Purpose of the Study:
- To investigate the dynamics of steering and obstacle avoidance in human navigation.
- To predict human routes through complex virtual scenes based on environmental cues.
Main Methods:
- Participants navigated a virtual environment towards a goal and around an obstacle.
- Observed steering behavior was modeled as a dynamical system.
- Angular acceleration was analyzed as a function of goal/obstacle angle and distance.
Main Results:
- Goals act as attractors, and obstacles as repellers of heading, with influence varying by distance.
- A dynamical system model accurately predicted steering behavior.
- Linear combination of goal and obstacle terms predicted route selection (left/right of obstacle).
Conclusions:
- Route selection in complex environments may emerge from online steering dynamics.
- Explicit path planning may not be necessary for navigation.
- The findings have implications for virtual reality and autonomous robot navigation.