Using bat-modelled sonar as a navigational tool in virtual environments

  • Authors:
  • Dean A. Waters;Husam H. Abulula

  • Affiliations:
  • Institute of Integrative and Comparative Biology, L.C. Miall Building, University of Leeds, Leeds LS2 9JT, UK;Institute of Integrative and Comparative Biology, L.C. Miall Building, University of Leeds, Leeds LS2 9JT, UK

  • Venue:
  • International Journal of Human-Computer Studies
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Bats are able to use active sonar as a mechanism for locating object in three dimensions and for generating spatial maps of their environments. Humans use passive sound cues to detect features of the space they occupy, as well as react to the spatial location of objects which generate sound. The system described in this paper allows free-ranging humans to locate a virtual sound location using active sonar. An emitted pulse, centred on the users head, serves as an intensity and time marker. The return pulse is rendered at the virtual target location and emitted after a time delay corresponding to the two-way path from sender to target and back again. The sonar system is modelled on those of bats, using ultrasonic frequency-modulated signals reflected from simple targets. The model uses the reflectivity characteristics of ultrasound, but the frequency and temporal structure used are scaled, with the speed of sound being set to 8.5ms^-^1 to bring the frequency range and temporal resolution within the capabilities of the human auditory system. Orientation with respect to the ensonified target is achieved by time-of-flight time delays to give target range, and binaural location information derived from interaural timing differences, interaural intensity differences, and head-related transfer functions. Subjects performed significantly better at a localization task when given temporal data based on echo delays with an outgoing reference pulse than without a reference pulse. Frequency-modulated signals sweeping from 1.5kHz-100Hz over 500ms provide the best localization cues, and users found them significantly easier to locate than continuous sounds.