| GPS can provide and has provided down to sub-centimeter accurate precise orbit determination (POD) for low Earth orbit (LEO) spacecraft. This is achieved using carrier-phase measurements, and for maximum accuracy, continuous tracking is needed to resolve the unknown constant carrier cycle ambiguities. Continuous GPS tracking can be difficult to obtain on small satellites, such as CubeSats, due to their stringent onboard power resources. This often leads to the GPS receiver power duty cycling during operations to allow for more on-board power to be used for other scientific instruments. This thesis investigates the sensitivity of GPS solution accuracy to duty cycled observations. In addition, it considers combining the duty cycled GPS observables with a simple dynamic model through an extended Kalman Filter (EKF) to maintain high accuracy POD despite duty cycling. Tests are conducted on simulated observables which are generated for 10 different satellite orbit configurations and varying amounts of GPS receiver duty cycle. Comparisons between, a simple GPS kinematic tracking and a two-body reduced dynamic approach are investigated, and the twobody reduced dynamic approach is shown to yield a 3D root mean square error (RMSE) that is significantly better than the kinematic approach, primarily during the initial solution convergence period. This thesis also investigates the effect of modeling the Earth's oblateness when integrating the state of the spacecraft in concert with a two-body model to propagate the error-covariance of the state. |
