The ANML Language, D. Smith & W. Cushing, to appear in: Proc. 9th
International Symposium on AI, Robotics, and Automation in Space
(iSAIRAS-08)
[pdf 96K]
Automatically Generated Heuristic Guidance for Europa2, S. Bernardini & D. Smith, to appear in: Proc. 9th International Symposium on AI, Robotics, and Automation in Space (iSAIRAS-08) [pdf 220K]
Describes the use of Monte Carlo
techniques for estimating cost in probabilstic planning.
A graphical depiction of some of the
different translations that have been done between different planning
formalisms.
A survey of recent techniques and results
for dealing with concurrency and time in probabilistic planning.
Developing Domain-Independent Search Control for Europa2, S. Bernardini & D. Smith, Proc. ICAPS-07 Workshop on Heuristics for Domain-independent Planning: Progress, Ideas, Limitations, Challenges [pdf 204K]
The Perils of Discrete Resource Models, W. Cushing & D. Smith, Proc. ICAPS-07 Workshop on International Planning Competition: Past, Present and Future [pdf 252K]
Preventing Unrecoverable Failures through Precautionary Planning, J. Foss, N. Onder & D. Smith, Proc. ICAPS-07 Workshop on Moving Planning and Scheduling Systems into the Real World [pdf 244K]
Describes a method for doing incremental contingency planning in the context of replanning.
Using Interaction to Compute Better Probability Estimates in Plan Graphs, D. Bryce & D. Smith, Proc. ICAPS-06 Workshop on Planning Under Uncertainty and Execution Control for Autonomous Systems, 2006. (Revised version - older version printed in proceedings.) [pdf 252K]
Describes method for keeping track of interaction and synergy between propositions in a Plan Graph to get better heuristic probability estimatesSequential Monte Carlo for Probabilistic Planning Reachability Heuristics, D. Bryce, S. Kambhampati, & D. Smith, Proc. 16th Intl. Conf. on Automated Planning & Scheduling (ICAPS-06), 2006. [pdf 204K]
Describes a Plan Graph heuristic method using Monte Carlo Simulation to guide search in probabilistic planningPlanning Graph Heuristics for Belief Space Search, D. Bryce, S. Kambhampati, & D. Smith, Journal of Artificial Intelligence Research 26, pages 35-99, 2006. [pdf 636K]
Develops and investigates heuristics for guiding belief space search using a Labelled Uncertainty GraphField Demonstration of Surface Human-Robotic Exploration Activity, L. Pedersen, W. Clancey, M. Sierhaus, N. Muscettola, D. Smith, D. Lees, K. Rajan, S. Ramakrishnan, P. Tompkins, A. Vera. T. Dayton. Proc. AAAI 2006 Spring Symposium: To Boldly Go Where No Human-Robot Team Has Gone Before, 2006. [pdf 1.2M]
Describes field demonstration of interactive commanding, planning and execution of tasks for human-robotic exploration
Oversubscription planning by searching in the space of plan strategiesMission Planning and Target Tracking for Autonomous Instrument Placement, L. Pedersen, D.Smith, M. Dean, R. Sargent, C. Kunz, D. Lees & S. Rajagopalan. Proc. 2005 IEEE Aerospace Conf., 2005. [pdf 1.7M]
Describes methods for planning and target tracking for rover insrument placementMultiple-Target Single Cycle Instrument Placement, L. Pedersen, M. Deans, D. Lees, S. Rajagopalan, and D. Smith, Proc. Eighth International Symposium on AI, Robotics, and Automation in Space (iSAIRAS-05), Munich, Sept, 2005. [pdf 1.5M]
Description of an integrated rover demonstration.
Describes a compact encoding of plan graphs for multiple possible worlds in a single labelled plan graphChoosing Objectives in Over-Subscription Planning, D. Smith. Proc. 14th Intl. Conf. on Automated Planning & Scheduling (ICAPS-04), 2004. [pdf 200K]
Describes a method for generating and solving an Orienteering Problem to choose the subset of goals for an oversubscription problem
Critique of condition/effect restrictions for durative actions in PDDL2.1The Next Challenges for AI Planning, D. Smith. Lecture given at the Planet International Summer School, Madonna di Campiglio, Italy, June 2003. [powerpoint 20M]
Presentation of six technical challenges for AI planning, motivated by NASA planning problemsOptimal Limited Contingency Planning, N. Meuleau & D. Smith. Proc. UAI-03 & ICAPS-03 Workshop on Planning under Uncertainty, Trento, Italy, June 2003. [pdf 160K]
Describes a layered POMDP approach to producing optimal plans having a limited number of conditional branches.Incremental Contingency Planning, R. Dearden, N. Meuleau, S. Ramakrishnan, D. Smith, & R. Washington. Proc. ICAPS-03 Workshop on Planning under Uncertainty, Trento, Italy, June 2003. Paper: [pdf 116K] Presentation: [powerpoint 2.3M]
Outlines an approach to contingency planning with uncertainty in action duration and continuous resources. Updated version of "Contingency Planning for Planetary Rovers".Interleaved Observation Execution and Rescheduling on Earth Observing Systems, L. Khatib, J. Frank, R. Morris, D. Smith and J. Dungan. Proc. ICAPS-03 Workshop on Plan Execution, Trento, Italy, June 2003. [pdf 184K]
Describes an incremental approach to onboard rescheduling of satellite observations.Integrated Demonstration of Instrument Placement, Robust Execution and Contingency Planning, L. Pedersen, M. Bualat, D. Lees, D. Smith, and R. Washington, Proc. Seventh International Symposium on AI, Robotics, and Automation in Space (iSAIRAS-03), Japan, May, 2003. [pdf 1.4M]
Description of an integrated rover demonstration.
Outlines an approach to contingency planning with uncertainty in action duration and continuous resources.Planning Under Continuous Time and Resource Uncertainty: A Challenge for AI, J. Bresina, R. Dearden, N. Meuleau, S. Ramakrishnan, D. Smith, and R. Washington. Proc. UAI-02, Edmonton, Alberta, July 2002. Paper: [pdf 1700K] AIPS-02 workshop presentation: [powerpoint 1400K]
A challenge paper describing planning problems with uncertainty in action duration and continuous resources .The Logic of Reachability, D. Smith and Ari Jónsson. Proc. Sixth International Conference on AI Planning and Scheduling (AIPS-02), Toulouse, France, April 2002. [pdf 116K]. AIPS-02 presentation: [powerpoint 1000K]
Formalization and extension of the notions of reachability and mutual exclusion to actions where preconditions and effects can hold over arbitrary intervals, and exogenous events can occur.Fragment-based Conformant Planning, J. Kurien, P. Nayak, and D. Smith. Proc. Sixth International Conference on AI Planning and Scheduling (AIPS-02), Toulouse, France, April 2002. [pdf 108K]. AIPS-02 presentation: [powerpoint 1800K]
Describes an incremental method of conformant planning that finds plans in one possible world and tries to extend them to work in subsequent possible worlds.
Description of system for scheduling observations on fleets of satellites.The Prospects for Real Planning, D. Smith. Invited talk at DX01: Twelfth Internationa Workshop on Principles of Diagnosis, Sansicario, Via Lattea, Italy, March 2001. [powerpoint 1000K]
An introduction to some of the important advances made in "Classical Planning" in the last few years. A written version of much of this material can be found in the paper Bridging the Gap Between Planning and Scheduling, listed below.
A critical overview of generative planning techniques for dealing with continuous time and resources. A written version of the material for this talk can be found in Section 4 of the paper Bridging the Gap Between Planning and Scheduling, listed below.Bridging the Gap Between Planning and Scheduling, D. Smith, J. Frank, and A. Jónsson, in Knowledge Engineering Review 15:1, 2000. [pdf 129K]
A critical overview of AI planning and scheduling techniques from a spacecraft applications point of view.
Develops a Graphplans-style algorithm that works for actions with arbitrary duration.Increasaed Flexibility and Robustness of Mars Rovers, J. Bresina, K. Golden, D. Smith, and R. Washington, Proc. Fifth International Symposium on AI, Robotics, and Automation in Space (iSAIRAS-99), June, 1999. [pdf 258K]
Description of the Contingent Rover Language (CRL), ground tools, and on-board executive developed and used for a Feb 99 Marsokhod rover field test.Autonomous Rovers for Mars Exploration, R. Washington, J. Bresina, K. Golden, D. Smith, C. Anderson and T. Smith, Proc. 1999 IEEE Aerospace Conference, Aspen, CO, 1999. [pdf 323K]
Overview and advocacy of an architecture for increasing the autonomy of future Mars rovers.
Technical note that describes a more compact way of building a plan graph. This incremental method avoids a great deal of duplicated work during graph expansion.Autonomous Rovers for Human Exploration of Mars, J. Bresina, G. Dorais, K. Golden, D. Smith and R. Washington, Proc. Mars Society Founding Convention, Boulder, CO, 1998. [pdf 191K]
Overview and advocacy of an architecture for increasing the autonomy of future Mars rovers.Extending Graphplan to Handle Uncertainty & Sensing Actions, D. Weld, C. Anderson and D. Smith, Proc. Sixteenth National Conference on Artificial Intelligence (AAAI-98), Madison, WI, 1998. [pdf 188K]
Describes a possible worlds approach to incorporating sensory actions and contingency branches into the Graphplan framework. We recommend you read the Conformant Graphplan paper first.Conformant Graphplan, D. Smith and D. Weld, Proc. Sixteenth National Conference on Artificial Intelligence (AAAI-98), Madison, WI, 1998. [pdf 44K]
Describes a possible worlds approach to doing conformant or "fail-safe" planning in the Graphplan framework.Conditional Effects in Graphplan, C. Anderson, D. Smith and D. Weld, Proc. Fourth International Conference on AI Planning Systems (AIPS-98), Pittsburgh, PA, 1998. [pdf 203K]
Describes a way of treating conditional effects and quantification (ADL operators) in Graphplan.
Presents a strategy for suspending and pruning recursive open conditions during POCL planning.Constraint Management Methodology for Conceptual Design Tradeoff Studies, S. Reddy, K. Fertig, and D. Smith, Proc. 1996 ASME Design Engineering Technical Conferences and Computers in Engineering Conference, Irvine, California, 1996. [pdf 247K]
Description of Design Sheet and use of the system for performing tradeoff studies on a thermal imaging performance model developed for the DARPA MADE program.
Describes a hybrid AI/Markov chain approach to evaluating the expected cost and probability of success for plans with conditional branches and loops.
Final report for Rockwell's phase 1 ARPA planning initiaive contract. Contains brief descriptions of a variety of work. My part contains brief descriptions of 1) conditional partial-order planning, 2) use of operator graphs to determine operator relevance, 3) a variable analysis technique for operator graphs, 4) a least comittment open condition ordering technique, 5) recursion analysis techniques, 6) threat postponement analysis, 7) threat resolution strategies, and 8) an operator graph mechanism for partial plan evaluation.
Unpublished addendum to the paper Threat Removal Strategies in Partial-Order Planning. Describes an additional delay strategy reported in the talk at AAAI 93.Threat-Removal Strategies for Partial-Order Planning, M. Peot and D. Smith, Proc. Eleventh National Conference on Artificial Intelligence (AAAI-93), Washington D.C., 1993. [pdf 50K]
Compares several different strategies for when to resolve threats in partial-order planning. Describes a strategy DSEP that provably dominates the strategy used in SNLP.Postponing Conflicts in Partial-Order Planning, D. Smith and M. Peot, Proc. Eleventh National Conference on Artificial Intelligence (AAAI-93), Washington D.C., 1993. [pdf 46K]
Describes a method for analyzing the operators, goal and initial conditions for a planning problem to determine which threats can be ignored during the planning process. An abbreviated version of this paper appeared in the AAAI Spring Symposium on the Foundations of Automated Planning: The Classical Approach and Beyond, Stanford, CA 1993.
Introduces and explains an algorithm for conditional partial-order planning.A Critical Look at Knoblock's Hierarchy Mechanism, D. Smith and M. Peot, Proc. First International Conference on AI Planning Systems (AIPS-92), College Park, Maryland, 1992. [pdf 13K]
Shows examples and explains why Knoblocks hierarchy mechanism can produce arbitrarily poor open condition orderings.Design Sheet: An Environment for Facilitating Flexible Trade Studies during Conceptual Design, M. Buckley, K. Fertig, and D. Smith, Proc. 1992 Aerospace Design Conference, American Institute of Aeronautics and Astronautics (AIAA), 1992. [pdf 206K]
Description of Design Sheet and the basic graph algorithms it uses to: 1) recognize when variables are determined, 2) decide what simultaneous systems of equations must be solved, and 3) determine the best iteration variables for numerical solution of simultaneous systems of nonlinear equations.
Reasoning about action I: a possible worlds approach, M. Ginsberg and D. Smith, Artificial Intelligence 35(2), pages 165--195, 1988. Reprinted in Readings in Nonmonotonic Reasoning, pages 433-463, Morgan Kaufmann, 1987. [pdf 245K]