Home > Papers from 2012 > Testing landmark vector models of robot homing

Testing landmark vector models of robot homing

Landmark vector methods are a subset of the insect-inspired class of visual homing methods. Following the original ALV, these models all involves storing vectors directed at specific parts of the panorama. In this paper, Yu et al investigate the performance of a set of landmark vector models.

Abstract: “The development of an autonomous navigating robot is a challenging task. Motivated by the performance of insects successfully returning to the nest, researchers have studied bio-inspired navigation algorithms for their potential use in mobile robots. In this paper, we analyze landmark-based approaches, especially Distance Estimated Landmark Vector (DELV), Average Correctional Vector and Average Landmark Vector methods, that use landmark vectors for visible environmental landmarks. We evaluated the homing performance of various landmark vector methods with surrounding landmarks under occlusion and found that the occluded or missing landmarks have a significant influence on the performance. We also developed a landmark vector algorithm with a visual compass that uses only retinal images without a reference compass. From our experimental results, we conclude that the DELV shows robust homing navigation performance with missing or occluded landmarks among landmark vector methods.”

Yu, S.-E., Lee, C. & Kim, D. 2012 Analyzing the effect of landmark vectors in homing navigation. Adaptive Behavior 20, 337-359

Advertisements
Categories: Papers from 2012
  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