Future NASA missions requiring spacecraft formation flying require an extremely high level of autonomy and robustness when compared to single spacecraft systems. This is especially true for formations with a large number of spacecraft, which naturally have a higher likelihood of collision, and those that are to be flown in deep space, which are located too far from the Earth to allow for direct ground-based control. Further, it is likely that the individual spacecraft will have a limited amount of resources for sensing and communication. This dissertation is devoted to the development of decentralized navigation and control algorithms for such systems. The algorithms developed efficiently utilize the limited sensing and communication resources at each spacecraft in order to maintain an accurate estimate of the formation state. Formation keeping is achieved through the calculation of a reference point which damps noise in the formation state estimates, of which the reference point is a function. In the absence of an intra-spacecraft communication subsystem, optimal sensor switching algorithms are developed which yield accurate formation state estimates. With communication, individual spacecraft state estimates are iteratively fused to form formation-optimal state estimates. Numerical simulations demonstrate the efficacy of these methods when compared, in terms of fuel usage and formation positioning error, to an ideal system where each spacecraft has comparatively high sensing and communication capability.