Home > Papers from 2015 > Rotational image matching without systematic scanning

Rotational image matching without systematic scanning

Over recent years, experimental and theoretical studies have promoted the idea that visual navigation can be undertaken if an agent simply moves in the most familiar direction. This is possible if the visual experience of an agent is shaped in adaptive ways during learning. Subsequent navigation can be driven by comparing views of the world when facing in a variety of directions against one’s set of memories  (rotational image matching). Whilst it can be shown this strategy works from an informational perspective, exactly how an agent should perform the matching, without intermittently scanning on the spot, is unclear. Kodzhabashev and Mangan create an algorithm that reduces the need for rotational scanning; Rotational movements are modulated by current visual familiarity such that paths are straight when aligned with familiar routes and when in an unfamiliar location scanning is up-regulated, something that is seen in ants. The algorithm is tested on a robot platform and in simulation.

Kodzhabashev, A., & Mangan, M. (2015). Route Following Without Scanning. In Biomimetic and Biohybrid Systems (pp. 199-210). Springer International Publishing.
Advertisements
Categories: Papers from 2015
  1. No comments yet.
  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s