This paper deals with an adaptive multi-robot task planning. When a desired goal state of the task environment is given, backward task planning from the goal state and forward search form the initial condition such as currently available robot and functions and displacement of objects etc., are interactively executed to find a feasible state transition. Nontrivial multi-robot cooperation formation is dynamically generated as a result of the task planning. Experimental results validate the proposed method.
Published in:
Robotic Intelligence In Informationally Structured Space (RiiSS), 2011 IEEE Workshop on
Date of Conference: 11-15 April 2011