A Real-Time Rover Executive Based On Model-Based Reactive Planning

摘要: This paper reports on the experimental verification of ability IDEA (Intelligent Distributed Execution Architecture) effectively operate at multiple levels abstraction in an autonomous control system. The basic hypothesis is that a large system can be structured as collection interacting agents, each organized around same fundamental structure. Two system-level agent and mission-level agent, are designed implemented to autonomously K9 rover real-time. evaluated scenario where must acquire images from specified set locations. agents responsible for enabling achieve its goals while monitoring execution safety recovering dangerous states when necessary. Experiments carried out both simulation physical rover, produced highly promising results.