Beschreibung
This book investigates the problem of competently accomplishing everyday manipulation activities for autonomous robots as a plan-based control problem. We propose TRANER (TRAnsformational PlanNER) as a plan transformation system for the optimization of everyday activities. TRANER advances the state-of-the-art in three important ways. First, it extends concurrent reactive plan representation to support the specification of robust and transformable plans. Second, it proposes a library of general and flexible plans for a household robot. Third, it provides a suite of general transformation rules that are capable of improving the course of activities in important and novel ways. Comprehensive and extensive experiments in a realistic simulation environment show that transformational planning enables robots to substantially improve their performance by tailoring their course of activity to their tasks, environments, and capabilities.
Autorenportrait
Armin Müller studied Computer Science at TechnischeUniversität München. His work on plan-based control of autonomousrobots has been presented at leading international conferences.He received his doctor\'s degree from TU München.
