Content area
In recent years, the rapid increase in the number of space objects has made space object catalog maintenance increasingly challenging, necessitating efficient observation strategies. Optical sensors play a crucial role in observing and tracking these objects, but optimizing their tasking remains a key problem. This work presents a practical implementation of the theoretical sensor tasking framework. A greedy optimizer is used to find the optimal viewing directions at every observation time. By generating computationally efficient, effective, and precise observation schedules, a sensor is able to execute observation tasks to allow for the successful observation of objects. Improvements on the sensor tasking framework are described by incorporating a set of optical visibility conditions that allow for more realistic observations. Observation campaigns were set up to validate the observation schedules using multiple sensors at different locations. The campaign results are shown and analyzed to demonstrate the efficiency and accuracy of the implemented sensor tasking algorithm. In real-world scenarios, the time a telescope takes to slew from one viewing direction to another is not constant. Therefore, a new formulation of the sensor tasking problem is introduced to account for the variable repositioning time of a sensor. Genetic algorithms are used to provide a stochastic optimization approach where candidate solutions evolve and mutate towards more optimal results. Theoretical results and comparisons between the genetic algorithm and a greedy approach to the variable repositioning time problem are shown. The genetic algorithm designed in this work provides a faster solution to the variable repositioning time problem as compared to the local optimization strategy of the greedy approach.