Author name cluster

Anthony Barrett

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

4 papers

2 author rows

ICAPS Conference 2004 Conference Paper

From Hybrid Systems to Universal Plans via Domain Compilation

Anthony Barrett

Whileuniversal plans tell a system how to reach a goal regardless of what state it is in, such plans can be too large to represent. Hybrid systems execute plans where each action is implemented to robustly produce effects if the world does not stray outside the action’s control envelope. This paper presents a middle ground between these two extremes that uses plans, but also enables much larger control envelopes using a real-time planner that finds optimal n step plans to achieve a set of goals if one exists.

Details

IROS Conference 2001 Conference Paper

Representation and execution of plan sequences for multi-agent systems

Paolo Pirjanian
Terrance L. Huntsberger
Anthony Barrett

Integration of planning and execution for single-agent systems has received considerable attention and numerous interesting approaches have been proposed. This is not the case for multi-agent systems, however. We describe one approach to representing joint team activities using a finite state machine augmented with synchronization primitives for orchestration of group activities. This representation encodes a sequence of activities that can be distributed across a team of robots and executed in a coordinated manner. We then show how such a representation can be mapped onto a behavior-based formalism for execution, as supported by a distributed, multi-robot control architecture, CAMPOUT. We demonstrate the proposed mechanisms within the context of a challenging task, where two rovers collectively carry an extended container over rough terrain.

Details

AIJ Journal 1994 Journal Article

Partial-order planning

Anthony Barrett
Daniel S. Weld

Although most people believe that planners that delay step-ordering decisions as long as possible are more efficient than those that manipulate totally ordered sequences of actions, this intuition has received little formal justification or empirical validation. In this paper we do both, characterizing the types of domains that offer performance differentiation and the features that distinguish the relative overhead of three planning algorithms. As expected, the partial-order (nonlinear) planner often has an advantage when confronted with problems in which the specific order of the plan steps is critical. We argue that the observed performance differences are best understood with an extension of Korf's taxonomy of subgoal collections. Each planner quickly solved problems whose subgoals were independent or trivially serializable, but problems with laboriously serializable or nonserializable subgoals were intractable for all planners. Since different plan representations induce distinct search spaces, the subgoals for a given problem may be trivially serializable for one planner, laboriously serializable for another, and nonserializable for a third. We contend that the partial-order representation yields superior performance because it more frequently results in trivial serializability.

Details DOI

AAAI Conference 1994 Conference Paper

Task-Decomposition via Plan Parsing

Anthony Barrett

Task-decomposition planners make use of schemata that define tasks in terms of partially ordered sets of tasks and primitive actions. Most existing taskdecomposition planners synthesize plans via a topdown approach, called taslc reduction, which uses schemata to replace tasks with networks of tasks and actions until only actions remain. In this paper we present a bottom-up plan pursing approach to task-decomposition. Instead of reducing tasks into actions, we use an incremental parsing algorithm to recognize which partial primitive plans match the schemata. In essence, our approach exploits the observation that schemata are a convenient means for reducing search. We compile the schemata into a declarative search control language (like that used in machine learning research), which rejects plan refinements that, cannot be parsed. We demonstrate that neither parsing nor reduction dominates the other on efficiency grounds and provide preliminary empirical results comparing the two. We note that our parsing approach allows convenient, comparison (and combination) of different search control technologies, generates minimal plans, and handles expressive languages (e. g. , universal quantification and conditional effects) with ease.

PDF Details