Mixed Logical and Probabilistic Reasoning for Planning and Explanation Generation in Robotics
Robots assisting humans in complex domains have to represent knowledge and reason at both the sensorimotor level and the social level. The architecture described in this paper couples the non-monotonic logical reasoning capabilities of a declarative language with probabilistic belief revision, enabling robots to represent and reason with qualitative and quantitative descriptions of knowledge and degrees of belief. Specifically, incomplete domain knowledge, including information that holds in all but a few exceptional situations, is represented as a Answer Set Prolog (ASP) program. The answer set obtained by solving this program is used for inference, planning, and for jointly explaining (a) unexpected action outcomes due to exogenous actions and (b) partial scene descriptions extracted from sensor input. For any given task, each action in the plan contained in the answer set is executed probabilistically. The subset of the domain relevant to the action is identified automatically, and observations extracted from sensor inputs perform incremental Bayesian updates to a belief distribution defined over this domain subset, with highly probable beliefs being committed to the ASP program. The architecture's capabilities are illustrated in simulation and on a mobile robot in the context of a robot waiter operating in the dining room of a restaurant.