Abstract
Missions involving multiple spacecraft have become of great interest in the last decade as they offer a number of scientific and engineering advantages. Though already largely adopted for communication, geo-location (GPS) and meteorology purposes, only recently this paradigm is showing its potential benefits for Earth Observation and Space Exploration. Multiple platforms are crucial in the context of global monitoring and disaster management. The Global Monitoring for Environment and Security - GMES or the Disaster Monitoring constellation are the first examples of this trend. From the mission planning point of view, the use of multiple platforms is opening new challenges to the automated planning and scheduling systems whose aim is gaining maximum value from the constellation by optimising the use of on-board resources and by coordinating the different spacecraft. Hence, new approaches are needed to handle this level of complexity. The main goal of this research is the construction of a ground-based automated planning and scheduling system for the imaging campaign of an Earth Observation constellation. The target mission is the Disaster Monitoring Constellation, which requires a system that is responsive to the asynchronous requests of different user groups with different priority levels. Multi agent systems represent a fruitful approach to model such a dynamic context. The novelty of this project is to apply nature-inspired techniques, such as stigmergy, to achieve optimisation and coordination. This mechanism offers high-level of adaptability and scalability allowing the system to find an efficient schedule at global level due to the collaboration of all the agents. A key novelty of this project is the development of a theoretical framework to model the self-organising long-term system’s behaviours. This model is able to describe the architecture as a dynamical system. It offers new insights which are the basis of a new algorithm which regulates the trade-off of exploration/exploitation via changes in the system’s stability. The theoretical model, as well as the algorithm, has been extended in order to include a coordination mechanism which is required by the multiple platform scenario. An empirical evaluation has been used to validate the system’s capabilities in optimisation, adaptability and scalability in the case of dynamic problems for single and multiple spacecraft. Lastly, the transferability of the system developed has been demonstrated to different contexts outside the Earth Observation field such as the ESA GENSO (Global Educational Network for Satellite Operations) network. This is a ground station network sharing similar requirements with the Earth Observation constellation scenario.