Autonomous adaptive exploration using realtime online spatiotemporal topic modeling

Abstract

The exploration of dangerous environments such as underwater coral reefs and shipwrecks is a difficult and potentially life-threatening task for humans, which naturally makes the use of an autonomous robotic system very appealing. This paper presents such an autonomous system, which is capable of autonomous exploration, and shows its use in a series of experiments to collect image data in challenging underwater marine environments. We present novel contributions on three fronts. First, we present an online topic-modeling-based technique to describe what is being observed using a lowdimensional semantic descriptor. This descriptor attempts to be invariant to observations of different corals belonging to the same species, or observations of similar types of rocks observed from different viewpoints. Second, we use the topic descriptor to compute the surprise score of the current observation. This is done by maintaining an online summary of observations thus far, and then computing the surprise score as the distance of the current observation from the summary in the topic space. Finally, we present a novel control strategy for an underwater robot that allows for intelligent traversal, hovering over surprising observations, and swimming quickly over previously seen corals and rocks.

Phytoplankton hotspot prediction with an unsupervised spatial community model

“…In topic models, Dirichlet distribution or Dirichlet process [14] priors can be used to control the sparseness of the taxonomic distribution representing a community, and the sparseness of the community distribution at a given location. Girdhar et al [15], [16] extended the standard topic model to account for spatial and temporal correlation of observations. The plankton community model we propose here is based on the Bayesian nonparametric spatio-temporal topic model (BNP-ROST) [16].…”

Towards a Human Factors Model for Underwater Robotics

“…sAUVs have the possibility to help overcome some of these challenges. Underwater exploration is a difficult and potentially life threatening operation for researchers [2]. Thus, some sAUVs have been developed by researchers to have the capability to help monitor marine environment health.…”

“…The videos were recorded between 2010 and 2013 during field trials [2,3]. 25 videos were analyzed with an average duration of 60 seconds each.…”

Anomaly detection in unstructured environments using Bayesian nonparametric scene modeling

“…Our proposed approach to dealing with the anomaly detection problem is to first use a Bayesian non-parametric scene understanding technique to build a model of the scene, *This work was supported by the Postdoctoral Scholar Program at the Woods Hole Oceanographic Institution, with funding provided by the Devonshire Foundation and the J. Seward Johnson Fund, and FQRTN Postdoctoral Fellowship. 1 and then using this model identify observations that are perplexing for the model. BNP topic modeling techniques have been successful in building semantic models of the data that automatically grow in complexity with data complexity.…”