• Classified by Topic • Classified by Publication Type • Sorted by Date • Sorted by First Author Last Name • Classified by Funding Source •
iCORPP: Interleaved commonsense reasoning and probabilistic planning on robots.
Shiqi
Zhang, Piyush Khandelwal, and Peter
Stone.
Robotics and Autonomous Systems, 2024.
Robot sequential decision-making in the real world is a challenge because itrequires the robots to simultaneously reason about the current world state anddynamics, while planning actions to accomplish complex tasks. On the one hand,declarative languages and reasoning algorithms support representing and reasoningwith commonsense knowledge. But these algorithms are not good at planning actionstoward maximizing cumulative reward over a long, unspecified horizon. On theother hand, probabilistic planning frameworks, such as Markov decision processes(MDPs) and partially observable MDPs (POMDPs), support planning to achievelong-term goals under uncertainty. But they are ill-equipped to represent orreason about knowledge that is not directly related to actions. In this article,we present an algorithm, called iCORPP, to simultaneously estimate the currentworld state, reason about world dynamics, and construct task-orientedcontrollers. In this process, robot decision-making problems are decomposed intotwo interdependent (smaller) subproblems that focus on reasoning to “understandthe world” and planning to “achieve the goal” respectively. The developedalgorithm has been implemented and evaluated both in simulation and on realrobots using everyday service tasks, such as indoor navigation, and dialogmanagement. Results show significant improvements in scalability, efficiency, andadaptiveness, compared to competitive baselines including handcrafted actionpolicies.
@Article{shiqi_ras2024,
author = {Shiqi Zhang and Piyush Khandelwal and Peter Stone},
title = {iCORPP: Interleaved commonsense reasoning and probabilistic planning on robots},
journal = {Robotics and Autonomous Systems},
year = {2024},
abstract = {Robot sequential decision-making in the real world is a challenge because it
requires the robots to simultaneously reason about the current world state and
dynamics, while planning actions to accomplish complex tasks. On the one hand,
declarative languages and reasoning algorithms support representing and reasoning
with commonsense knowledge. But these algorithms are not good at planning actions
toward maximizing cumulative reward over a long, unspecified horizon. On the
other hand, probabilistic planning frameworks, such as Markov decision processes
(MDPs) and partially observable MDPs (POMDPs), support planning to achieve
long-term goals under uncertainty. But they are ill-equipped to represent or
reason about knowledge that is not directly related to actions. In this article,
we present an algorithm, called iCORPP, to simultaneously estimate the current
world state, reason about world dynamics, and construct task-oriented
controllers. In this process, robot decision-making problems are decomposed into
two interdependent (smaller) subproblems that focus on reasoning to “understand
the world†and planning to “achieve the goal†respectively. The developed
algorithm has been implemented and evaluated both in simulation and on real
robots using everyday service tasks, such as indoor navigation, and dialog
management. Results show significant improvements in scalability, efficiency, and
adaptiveness, compared to competitive baselines including handcrafted action
policies.
},
}
Generated by bib2html.pl (written by Patrick Riley ) on Tue Nov 19, 2024 10:24:38