Switching behavior in moving obstacle avoidance

Abstract

Purpose: Previous work by Warren, Di, & Fajen (VSS, 2003) investigated the paths people take to avoid a moving obstacle en route to a goal. These results were used to extend a dynamical model for stationary and moving obstacle avoidance (Fajen & Warren, JEP:HPP, 2003). The current study tests the model for the critical conditions in which participants switch from passing in front of a moving obstacle to passing behind it. We then generalize the model to environments containing both moving and stationary obstacles. Methods: Studies are conducted in the Virtual Environment Navigation Lab, a 12m × 12m space allowing for free movement in a virtual world. Participants wear a head-mounted display (HMD) with a 60° (H) × 40° (V) field of view of a virtual environment consisting of a textured ground plane. They are instructed to walk towards a stationary goal (blue post) while avoiding a moving obstacle (red post). Head position and orientation are recorded using a hybrid inertial/ultrasonic tracking system that samples their movements at 60 Hz, and updates the display with a 50 ms latency. Design: To create critical switching conditions, the speed and direction of the moving obstacle are manipulated. In Experiment 1, the obstacle moves on a 70°, 90° (frontal plane), or 110° trajectory, crossed with speeds of 0.4, 0.6, and 0.8 m/s. Experiment 2 adds a stationary obstacle near the participant's observed path. Implications: Results from these studies test and refine the obstacle avoidance components of the steering dynamics model. Related studies test the component for moving target interception (Owens & Warren, VSS 2004, 2005; Bruggeman & Warren, VSS 2005), with the aim of integrating these components into a general model of locomotor behavior in complex environments.