Mixed Logical Inference and Probabilistic Planning for Robots in Unreliable Worlds

Abstract

Deployment of robots in practical domains poses key knowledge representation and reasoning challenges. Robots need to represent and reason with incomplete domain knowledge, acquiring and using sensor inputs based on need and availability. This paper presents an architecture that exploits the complementary strengths of declarative programming and probabilistic graphical models as a step toward addressing these challenges. Answer Set Prolog (ASP), a declarative language, is used to represent, and perform inference with, incomplete domain knowledge, including default information that holds in all but a few exceptional situations. A hierarchy of partially observable Markov decision processes (POMDPs) probabilistically models the uncertainty in sensor input processing and navigation. Nonmonotonic logical inference in ASP is used to generate a multinomial prior for probabilistic state estimation with the hierarchy of POMDPs. It is also used with historical data to construct a beta (meta) density model of priors for metareasoning and early termination of trials when appropriate. Robots equipped with this architecture automatically tailor sensor input processing and navigation to tasks at hand, revising existing knowledge using information extracted from sensor inputs. The architecture is empirically evaluated in simulation and on a mobile robot visually localizing objects in indoor domains.

Towards an Explanation Generation System for Robots: Analysis and Recommendations

“…For instance, we are exploring the introduction of the probabilistic description of knowledge and uncertainty in conjunction with the logical reasoning. We hope to build on recent work that has been reported in this direction in the general context of using ASP and probabilistic planning with robots [33,38]. Furthermore, we are exploring the extension of KRASP and UMBRA to explicitly model communication between agents and to use this model for generating explanations.…”

“…ASP also supports reasoning in large knowledge bases [36] and reasoning with quantifiers. These capabilities have led to widespread use of ASP-based architectures in robotics by an international community of researchers [29,37,38]. In the remainder of this paper, any mention of KRASP refers to the use of these capabilities of ASP; where appropriate, we highlight the differences between KRASP and a standard ASP formulation.…”

“…Recent work in this direction has been reported in the context of using ASP and probabilistic graphical models for planning and diagnostics in robotics [33,38,41]. A sophisticated explanation system could extend KRASP-like representations to hypothesize and evaluate different possible worlds [13,14], but incrementally guide the reasoning using well-designed heuristics, to generate explanatory structures that tightly couple probabilistic and logic-based representations of knowledge.…”

Service robot planning via solving constraint satisfaction problem

“…Answer set programming (ASP) is another popular approach used for planning [12,13]. It is a declarative language that is suitable for knowledge representation and non-monotonic reasoning.…”

“…It is a declarative language that is suitable for knowledge representation and non-monotonic reasoning. [12] integrated ASP with cost learning to improve the performance of the planner for robot planning, while [13] combines with MDP to endow it with the capability to handle uncertainty. At the moment, for complex sequence of services, these methods require heavy computational load and long planning time.…”