Tracking Sensor Network Schedule Optimization for Space Surveillance

Thousands of satellites and hundreds of thousands of pieces of space debris are constantly orbiting our planet. An important part of preventing collisions between these objects is space surveillance. In other words, tracking of all these objects, to be able to predict collisions before they happen. One way of tracking these objects is by using an optical ground station, also referred to as an optical sensor, or a telescope. To be able to maintain an accurate estimate of the orbits of these objects, they have to be reobserved frequently. At the same time, it is only possible to observe them during certain time windows, when the observation geometry is correct and the weather is favorable. It is therefore very challenging to make a schedule for when to observe each object, especially if a whole network of sensors needs to be scheduled. In this thesis, an automated scheduler has been developed which is capable of autonomously producing a schedule for a network of optical sensors. Four different scheduling algorithms have been developed to enable schedule optimization for a wide range of different situations, objectives and requirements. Test results show that a near-optimum schedule can be produced in a fraction of the time required by an exact algorithm. The scheduler is also capable of dynamic rescheduling and optimization with respect to weather forecast.