Time-Constrained Exploration Using Toposemantic Spatial Models: A Reproducible Approach to Measurable Robotics

Abstract

As robots become ubiquitous in our everyday lives, they are more and more often assigned complex tasks involving multiple, simultaneous objectives. Moreover, efficiency, envisioned as the ability to complete more tasks in a given amount of time, is becoming increasingly important. At the same time, spurred by progress in machine learning, there is a tendency to explore novel designs in which robots rely more on visual sensors and less on traditional sensors like range finders.