Home alone: Autonomous extension and correction of spatial representations

Abstract

Abstract-In this paper we present an account of the problems faced by a mobile robot given an incomplete tour of an unknown environment, and introduce a collection of techniques which can generate successful behaviour even in the presence of such problems. Underlying our approach is the principle that an autonomous system must be motivated to act to gather new knowledge, and to validate and correct existing knowledge. This principle is embodied in Dora, a mobile robot which features the aforementioned techniques: shared representations, non-monotonic reasoning, and goal generation and management. To demonstrate how well this collection of techniques work in real-world situations we present a comprehensive analysis of the Dora system's performance over multiple tours in an indoor environment. In this analysis Dora successfully completed 18 of 21 attempted runs, with all but 3 of these successes requiring one or more of the integrated techniques to recover from problems.

Robot task planning and explanation in open and uncertain worlds

“…The knowledge predicate K(v) indicates that the value of v is known. 14 The assumptive predicate A(v, x) states that the planner has made the assumption 15 that v has the value x. Knowledge predicates obey a few rules that are enforced by the planning system: Definition 1.…”

“…This is supported in our system. 15 Our use of "assumption" differs somewhat from its everyday use in that we can assume a fact even if we know it to be true.…”

“…The entire architecture is an instantiation of the CoSy Architecture Schema (CAS) [6], implemented with the CAS Toolkit (CAST) [16]. This is a component-based, event-driven integration framework that has been used to develop multiple robot systems [45,15,13]. CAST enables development of loosely coupled, component-based systems governed by the exchange of information and knowledge via multiple working memories.…”

“…This topological map consists of nodes called places (corresponding one to one with the places in the relational map), connected through edges representing traversable routes. The underlying simultaneous localization and mapping techniques and obstacle perception methods are based on work by Hawes et al [15], Pronobis et al [37] and Zender et al [55]. For the purpose of this paper, it is enough to know that these techniques allow the robot to be well localized and to build a 3D metric map using information taken from 2D laser scans, 3D point clouds from a depth camera, and odometry data.…”

A system for interactive learning in dialogue with a tutor

“…This drive leads George to be curious about its world. We have previously shown the benefits of a motivation framework featuring such a drive in a mobile robot [17] and are now exploring its use in learning and dialogue. The knowledge extension drive has three associated goal generators.…”