Abstract
In this thesis we present two related aspects for studying the performance of an Earth Observation service provided by small satellites. The first part of the thesis explores the problem of predicting precise timings for when a LEO satellite will be above a given target on the ground. We present a new algorithm for LEO satellite orbit prediction - FPSCA (Fast Prediction of Satellite Closest Approach) which is suitable for satellite nadir tracking with respect to specific ground targets for imaging and communications. We demonstrate that this method is at least 220 times faster than the conventional approach. The new algorithm not only takes account of secular perturbations but also atmospheric drag and periodic variations along the orbit. The algorithm has been designed to run on the satellite so that the satellite can continuously determine an imaging schedule with asynchronous requests from multiple ground users. It also provides a capability for long-term orbit prediction to high level of accuracy. The second part of this thesis introduces the application of queueing theory to model the behaviour of small satellite providing a multi-user imaging service, of user defined ground target. This service has been analysed in terms of two queues - an image capture queue and an image download queue. As an extension of M/M/1 queue, we develop a new model which can be used to successfully analyse the image capture queue with a peak probability for a steady state number of requests to always reside in the queue. We also generalise M/M/1 queue to M/G/1 where the general output distribution is expressed in terms of orthogonal basis functions, the first of which degenerates to M/M/1. We prove that such queues are invertible, so that given the probability distribution of queue lengths we can solve for the service rate distribution required. Finally we show that the image download queue can be modelled in terms of a batch queue. Key words: orbit determination, imaging prediction, epicycle motion, small satellite, low Earth orbit, rise and set time, imaging service, queueing theory.