The ability to perform autonomous in-space inspection and servicing activities is of significant importance to the future space program. The University of Maryland has developed a pair of free-flying robotic platforms that operate in a neutral buoyancy environment to support such space-based activities. Numerous tests have been performed with this platform, including demonstration of a robust closed-loop controller for the rotational degrees of freedom. This rotational controller has been observed to substantially reduce operator workload through stable maintenance and tracking of attitude commands. However, in previous work the operator has directly controlled translational motion due to the absence of an inertial navigation system such as GPS in the underwater environment. Recently, a Visual Positioning System (VPS) has been developed for the neutral buoyancy environment that provides translational position and velocity estimates of the free-flying inspection vehicle. This paper describes the development and testing of a six degree-offreedom control system for this inspection vehicle and analyzes its performance during a variety of stationkeeping, pilot-assisted, and autonomous trajectory-tracking tasks.

Closed-loop attitude steering is a concept for implementing an attitude trajectory by using a conventional quaternion error feedback controller to track the time-varying attitude reference, rather than to simply regulate to a desired orientation. This is done by sampling the reference input and executing the maneuver as a sequence of closely spaced regulating commands that are read out from the spacecraft’s command buffer. The idea has been employed in practice to perform zero-propellant maneuvers on the International Space Station and minimum-time maneuvers on NASA’s TRACE space telescope as well as NASA’s Lunar Reconnaissance Orbiter (LRO). A challenge for operational implementation of the idea is the limited capacity of a space vehicle’s command storage buffer, which is normally not designed with attitude tracking in mind. One approach to mitigate the problem is to downsample-and-hold the attitude commands so that the attitude control system (ACS) regulates to a series of waypoints. This article explores the waypoint following dynamics of a quaternion error feedback control law for such an approach. It is shown that downsample-and-hold induces a ripple between downsamples that causes the satellite angular rate to significantly overshoot the desired limit. Analysis in the z-domain is carried out in order to understand the phenomenon. An interpolating Chebyshev-type filter is proposed that allows the desired attitude trajectory to alternatively be encoded in terms of a small set of filter coefficients. Using the interpolating filter, the continuous-time reference trajectory can be reconstructed and issued at the ACS rate but with significantly reduced memory requirements. The ACS of the LRO is used as an example to illustrate the behavior of a practical ACS.

No abstract

Closed-loop attitude steering can be used to implement a non-standard attitude maneuver by using a conventional attitude control system to track a non-standard attitude profile. The idea has been employed to perform zero-propellant maneuvers on the International Space Station and minimum time maneuvers on NASA's TRACE space telescope. A challenge for operational implementation of the idea is the finite capacity of a space vehicle's command storage buffer. One approach to mitigate the problem is to downsample-and-hold the attitude commands as a set of waypoints for the attitude control system to follow. In this paper, we explore the waypoint following dynamics of a quaternion error feedback control law for downsample-and-hold. It is shown that downsample-and-hold induces a ripple between downsamples that causes the satellite angular rate to significantly overshoot the desired limit. Analysis in the z-domain is carried out in order to understand the phenomenon. An interpolating Chebyshev-type filter is proposed that allows attitude commands to be encoded in terms of a set of filter coefficients. Using the interpolating filter, commands can be issued at the ACS rate but with significantly reduced memory requirements. The attitude control system of NASA's Lunar Reconnaissance Orbiter is used as an example to illustrate the behavior of a practical attitude control system.