Activity-based mission planning and plan management for autonomous vehicles

Abstract

This paper presents an activity-based mission planning and plan management framework that has been developed at the Charles Stark Draper Laboratory. The main contributions of this paper are definition of this activity-based implementation and comparison with other planning implementation approaches (e.g., behavior-based); and explanation of how planning, execution, monitoring, and replanning are implemented within this activity-based approach. One of the main benefits of this approach is the ability to separate the mission planning and plan management algorithms from activity specific algorithms and from mission and vehicle specific information. This separation results in an implementation that is highly portable both between missions and vehicles. This framework has been implemented and demonstrated in high-fidelity autonomous land and underwater vehicle simulations, is planned to be implemented on ARPA's UUV in the near future, and is being considered for applications involving other underwater vehicles, autonomous land rovers, and the space shuttle.