Robot Exploration Using Knowledge of Inaccurate Floor Plans

Abstract

In this paper, we propose a method that, given a partial grid map of an indoor environment built by an autonomous mobile robot, estimates the amount of the explored area represented in the map, as well as whether the uncovered part is still worth being explored or not. Our method is based on a deep convolutional neural network trained on data from partially explored environments with annotations derived from the knowledge of the entire map (which is not available when the network is used for inference). We show how such a network can be used to define a stopping criterion to terminate the exploration process when it is no longer adding relevant details about the environment to the map, saving, on average, 40% of the total exploration time with respect to covering all the area of the environment.

Exploration and Mapping with Groups of Robots: Recent Trends

“…Current work has incorporated a priori information in the exploration strategy to better inform the selection of the next locations to visit. Such information has been encoded in different forms: topo-metric graphs, e.g., provided by a user [56], and building floor plans [57,58]. Semantic information-i.e., spatial concepts given by humans, such as "corridors" and "offices"-could be also provided by a user so that the exploration strategy, together with a semantic mapping system, can focus to specific areas of interest and determine the number of robots to be assigned to a location [59].…”

Deep-Learning-Aided Path Planning and Map Construction for Expediting Indoor Mapping

“…Moreover, invariably the entire image is used in order to map only a local missed patch of it [2], [11], [14]. Another works on navigation over 2D maps, looks for an optimal local information gain actions [15], local matching topologies [16], navigating using reinforcement learning techniques [17], [18], or exploiting existing plans of the building [19]. Conversely, in the present work a prediction of the complete image is performed, given relatively few observations.…”