Autonomous power management system for a small satellite

An argument that a fuzzy system design methodology is better suited for the design of autonomous power systems than the currently used classical expert system methodology is presented. An architecture that incorporates fuzzy systems is developed for designing autonomous power management systems. The architecture was designed to meet both present and projected future technological requirements for satellite autonomy. Systems with varying degrees of autonomy can be implemented by varying the number of modules and their complexity. Detailed design, based on the proposed architecture, is given of a system with Level 5 autonomy. This system was implemented on a low cost 8-bit microprocessor and required less than 6 kBytes of memory. The power management system implemented was shown to perform the required autonomy functions through experiments in which it was required to follow a sequence of commands given in the form of a load switching schedule. After the power management system passed the command execution test, faults were introduced in the load to test its fault tolerance capabilities. The system was able to detect, identify and take appropriate action whenever an anomaly was introduced.