The LQG (linear-quadratic-Gaussian) controllers, designed for the NASA Deep Space Network antennas, have small tracking errors and are resistant to wind disturbances. However, they induce limit cycling during antenna-slewing operations and target acquisition. The cycling is caused by the violation of the antenna's rate and acceleration limits. This problem can be avoided by introduction of a command that does not exceed the limits. A command preprocessor, presented in this paper, generates a command that is equal to the original command if the latter does not exceed the limits, and that varies with the maximum (or minimum) allowable rate and acceleration if the limits are met or exceeded. The preprocessor is comparatively simple, since it requires the knowledge of the command at only the current and the previous instants of time, while the other known preprocessor [3] requires the knowledge of the terminal state and the acquisition time. Analysis and field tests of the preprocessor are presented in this article. The dynamics of the preprocessor itself, and of the antenna with the preprocessor, are illustrated with the antenna's measured responses to typical commands.
