An autonomous multi-sensor satellite system based on multi-agent blackboard model

Abstract

Traditional Earth observation satellite cannot work well in terms of emergencies, environmental uncertainties and scientific events discovery. Therefore, it is of significance to study the new generation of autonomous Earth observation satellite. In order to develop an autonomous satellite system with distributed and coordinated functions, this paper proposes an autonomous satellite system based on distributed multi-agent blackboard model. Multiple agents including functions of pre-processing, planning, scheduling and execution are designed. Agents share information and communicate through a blackboard which stores the task sequence, the action sequence and the satellite status. An adaptive rule-based heuristic scheduling algorithm and a forward search planning algorithm are proposed. The simulation experiments and computational results prove that the system can deal with scientific events discovery, satellite faults, cloud obscuration and emergencies without human intervention, which can greatly enhance the efficiency and reliability of Earth observation satellites. The validity of the proposed model and algorithm is proved.