A preliminary investigation of decentralized control far satellite formations

A distributed satellite formation consisting of n nodes interconnected via a communications network could be controlled using a decentralized controller framework that operates in parallel over the network. For such problems, a solution that minimizes data transmission requirements, in the context of linear-quadratic control theory, was given previously by Speyer. An investigation of the feasibility of such an approach to satellite formation flying is discussed in the context of efforts sponsored by NASA and the Air Force to develop architectures, strategies, and control approaches for various proposed distributed satellite missions. Among the issues under investigation are the effects of command and data handling system and communications channel noise and latency. For example, it is likely that the measurement devices and/or their associated command and data handling systems may provide asynchronous and/or non-simultaneous measurement data. It is also probable that transmission of the data between nodes could be delayed one or more sampling periods and/or interrupted for extended intervals. Furthermore, since it is not possible to create a noise-free communications channel, errors in the transmitted data will be introduced at the level of the network, in addition to errors introduced at the measurement and state transition levels. An additional difficulty is that, in general, nonlinearities in the dynamics and measurements of the distributed satellite control problem require the use of ad-hoc procedures such as the extended Kalman filter. Therefore, a simulation methodology is adopted for the study of the aforementioned issues in this work.