LaLaLoc: Latent Layout Localisation in Dynamic, Unvisited Environments
This addresses the challenge of robust localization for robotics or AR in changing domestic settings, representing a novel approach rather than an incremental improvement.
The paper tackles the problem of localizing in dynamic, unvisited environments without prior visitation, achieving accurate pose estimation with a localization error of 8.3cm using only a floor plan as prior.
We present LaLaLoc to localise in environments without the need for prior visitation, and in a manner that is robust to large changes in scene appearance, such as a full rearrangement of furniture. Specifically, LaLaLoc performs localisation through latent representations of room layout. LaLaLoc learns a rich embedding space shared between RGB panoramas and layouts inferred from a known floor plan that encodes the structural similarity between locations. Further, LaLaLoc introduces direct, cross-modal pose optimisation in its latent space. Thus, LaLaLoc enables fine-grained pose estimation in a scene without the need for prior visitation, as well as being robust to dynamics, such as a change in furniture configuration. We show that in a domestic environment LaLaLoc is able to accurately localise a single RGB panorama image to within 8.3cm, given only a floor plan as a prior.