Incorporating decision-theoretic planning in a robot architecture (original) (raw)
Related papers
Decision-Theoretic Planning in the Graphplan Framework
We have developed a decision-theoretic planner based upon the Graphplan planning algorithm, DT-Graphplan. DT-Graphplan reasons about probabilities, costs, and rewards at a propositional level, reconstructing limited state information. We are applying the planner to our robot task architecture to function on a miniature golf domain. By incorporating decision theory into planning, we seek to reduce the representational gap between behavior-based robotic controllers and constraint-based symbolic planners.
Planning Under Uncertainty Through Goal-Driven Action Selection
Lecture Notes in Computer Science, 2018
Online planning in domains with uncertainty and partial observability conveys a series of performance challenges: agents must obtain information about the environment, quickly select actions with high reward prospects and avoid very expensive mistakes, while interleaving planning and execution in highly variable and uncertain domains. In order to reduce the amount of mistakes and help an agent focus on directly relevant actions, we propose a goal-driven, action selection method for planning in (PO)MDP's. This method introduces a reward bonus and a rollout policy for MCTS planners, both of which depend almost exclusively on a clear specification of the goal and produced promising results when planning in large domains of interest to cognitive and mobile robotics.
1996
This paper shows how we can combine logical representations of actions and decision theory in such a manner that seems natural for both. In particular we assume an axiomatization of the domain in terms of situation calculus, using what is essentially Reiter's solution to the frame problem, in terms of the completion of the axioms defining the state change. Uncertainty is handled in terms of the independent choice logic, which allows for independent choices and a logic program that gives the consequences of the choices. As part of the consequences are a specification of the utility of (final) states. The robot adopts robot plans, similar to the GOLOG programming language. Within this logic, we can define the expected utility of a conditional plan, based on the axiomatization of the actions, the uncertainty and the utility. The 'planning' problem is to find the plan with the highest expected utility. This is related to recent structured representations for POMDPs; here we use stochastic situation calculus rules to specify the state transition function and the reward/value function. Finally we show that with stochastic frame axioms, actions representations in probabilistic STRIPS are exponentially larger than using the representation proposed here.
Using a contingency planner within a robot architecture
planiart.usherbrooke.ca
To achieve a given goal, a mobile robot must plan by predicting the possible evolution of the environment and the possible consequences of its actions. The use of actuators and sensors with limited precision and the presence of exogenous agents in the environment leads to nondeterministic predictions. However, most planbased robotic frameworks ignore this nondeterminism at the planning time, by producing only deterministic plans, and replanning whenever the outcome of actions or the the environment's dynamics stray away from the assumed ones. The main motivation behind this choice has been the longstanding lack of effective planners handling nondeterminism; but recent advances in this area make it possible, and advisable, to exploit such systems to implement more robust planbased robot behaviors. In this paper, we experiment the integration of a state-of-the art contingency planner in a robotic architecture, and discuss how such an integration improves the degree of flexibility and robustness of plan-based robot behaviors compared to the use of a deterministic planner.
A Constraint-Based Control Architecture for Acting and Reasoning in Autonomous Robots
1995
In this paper we will address several architectural decisions in defining a software control architecture for mobile robots. Our system is a collection of control primitives that enables the development of simulations or control algorithms for autonomous agents. Its computational capabilities are determined by an object-oriented constraint-based architecture. We discuss how high level knowledge, skills, goal-driven and reactive behavior are integrated within such an architecture. Our goal is to design a framework that enables the merging of classic and reactive implementation ideas. We will show, that each such type of control can be implemented in our system. The issues of task decomposition and granularity are given special attention, as they lie at the basis of our architecture. We discuss two learning methods supported by our system. The first is based on environment exploration, while the second copes with skill acquisition. Our robot, CYCLOPS , is a LEGO mini-robot based on th...
Rational Handling of Multiple Goals for Mobile Robots
Artificial Intelligence Planning Systems, 1992
The mobile robot planning domain is dynamic, with goals becoming active asynchronously. In order to successfully operate in this environment, a robot must be able to interrupt and reformulate its plan of action on-the-fly. This report investigates a method for incorporating the accomplishment of a new goal into a partially executed plan. A decision theoretic approach using net present value as the decision criterion serves as the basis for determining goal ordering dynamically. The appropriateness of net present value over other criteria is argued. The approach has been implemented on a robot operating in an office setting. Examples from this domain and a planetary exploration domain are used to show the advantages of the approach with respect to fixed priority and heuristic approaches.
Mixing Non-Monotonic Logical Reasoning and Probabilistic Planning for Robots
2015
This paper describes an architecture that combines the complementary strengths of probabilistic graphical models and declarative programming to represent and reason with qualitative and quantitative descriptions of domain knowledge and uncertainty. An action language is used for the architecture’s low-level (LL) and high-level (HL) system descriptions, and the HL definition of recorded history is expanded to allow prioritized defaults. For any given objective, each action in the plan created in the HL using non-monotonic logical reasoning is executed probabilistically in the LL, refining the HL description to identify the relevant sorts, fluents and actions, and adding the corresponding action outcomes to the HL history. The HL and LL domain representations are translated into an Answer Set Prolog (ASP) program and a partially observable Markov decision process (POMDP) respectively. ASP-based inference provides a multinomial prior for POMDP state estimation, and populates a Beta den...
Robot task planning and explanation in open and uncertain worlds
A long-standing goal of AI is to enable robots to plan in the face of uncertain and incomplete information, and to handle task failure intelligently. This paper shows how to achieve this. There are two central ideas. The first idea is to organize the robot's knowledge into three layers: instance knowledge at the bottom, commonsense knowledge above that, and diagnostic knowledge on top. Knowledge in a layer above can be used to modify knowledge in the layer(s) below. The second idea is that the robot should represent not just how its actions change the world, but also what it knows or believes. There are two types of knowledge effects the robot's actions can have: epistemic effects (I believe X because I saw it) and assumptions (I'll assume X to be true). By combining the knowledge layers with the models of knowledge effects, we can simultaneously solve several problems in robotics: (i) task planning and execution under uncertainty; (ii) task planning and execution in open worlds; (iii) explaining task failure; (iv) verifying those explanations. The paper describes how the ideas are implemented in a three-layer architecture on a mobile robot platform. The robot implementation was evaluated in five different experiments on object search, mapping, and room categorization.
Decision-making for a robotic architecture
This paper presents an overview of the intelligent decision-making capabilities of the CLARAty robotic architecture for autonomy. CLARAty is a two layered architecture where the top Decision Layer contains techniques for autonomously creating a plan of robot commands and the bottom Functional Layer provides standard robot capabilities that interface to system hardware. This paper focuses on the Decision Layer organization and capabilities. Specifically, the Decision Layer provides a framework for utilizing AI planning and executive techniques, which provide onboard, autonomous command generation and replanning for planetary rovers. The Decision Layer also provides a flexible interface to the Functional Layer, which can be tailored based on user preferences and domain features. This architecture is currently being tested on several JPL rovers.
Robot planning in the space of feasible actions: Two examples
1996
Several researchers in robotics and artijicial intelligence have found that the commonly used method ofplanning in a state (conjiguration) space is intractable in certain domains. This may be because the C-space has very high dimensionality, the "C-space obstacles" are too diflcult to compute, or; because a mapping between desired states and actions is not straightforward. Instead of using an inverse model that relates a desired state to an action to be executed by a robot, we have used a methodology that selects between the feasible actions that a robot might execute, in effect, circumventing many of the problems faced by configuration space planners. In this paper we discuss the implications of such a method and present two examples of working systems that employ this methodology. One system drives an autonomous crosscountry vehicle while the other controls a robotic excavator performing a trenching operation.