| Abstract
|
We present a control mechanism for selecti … We present a control mechanism for selecting/generating the current goal of planning/acting agents in unpredictible environments. This mechanism enables such an agent to generate and revise goals obtained by subgoal regression or by reaction to current events. In order to reach the agent's adaptibility to different environments/applications, we assume that every possible goal can be evaluated by this mechanism. A practical instantiation of this mechanism is used for controlling a mobile robot in indoor environments for object delivery and surveillance scenarios. Experimental results show dependencies between control and envirtonment characteristics. control and envirtonment characteristics.
|
| Author
|
Philippe Morignot +,
Barbara Hayes-Roth +
|
| Bibtype
|
techreport +
|
| Institution
|
Knowledge Systems, AI Laboratory +
|
| Key
|
KSL-95-75 +
|
| Modification dateThis property is a special property in this wiki.
|
1 May 2009 14:03:47 +
|
| Month
|
October +
|
| Number
|
KSL-95-75 +
|
| Tag
|
Computer science +
|
| Title
|
Goal Generation and Revision for Planning Agents in Unpredictable Environments +
|
| Tr id
|
KSL-95-75 +
|
| Year
|
1995 +
|
| Categories |
Technical Report,
Publication,
KSL Technical Report
|
