Quarterly Progress Report on Contract N00014-91-J-1577 (Yale University)
Abstract
We have designing and implementing a representation scheme and the basic planning algorithms for handling declarative goals in a reactive planner. The theory underlying the planning algorithms is based on three main paradigms: transformational planning, decision theory, and time-dependent computations. Knowledge about goals and the robot control language is accessed through transformation rules that define semantic relationships between constructs in the plan representation language and associations between goals and canned plans. The computational theory deals with uncertainty by applying estimation methods to determine the robustness of plans. The tradeoffs between planning and acting are weighed by applying time-dependent algorithms for testing the applicability and utility of transformation rules with respect to current situation and the preferences of the robot. Several features of the planar make it attractive for robot planning. The time-dependent planning algorithms guarantee that the planar proposes executable plans for any allocation of computation time. The approach is compatible with a general transformational framework. And, finally, decisions are controlled by explicitly modelled probabilities and utilities.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1992
- Accession Number
- ADA248832
Entities
People
- Drew McDermott
Organizations
- Yale University