Application of Nonequidistant Fractional-Order Accumulation Model on Trajectory Prediction of Space Manipulator

Abstract

To predict the trajectory of a motion target floating in an outer space, a nonequidistant fractional-order accumulation (NEFA) model of a visual-based space manipulator is presented in this paper. This model is a modified extrapolate method of trajectory prediction. First, the initial discrete point set of a motion trajectory is real-time accumulated to a fractional-order differential model. Second, a follow trajectory is predicted both by the fractional-order differential model and a dynamic data refresh strategy. A novel measuring machine of a space manipulator is designed to orbital simulation and trajectory predictive test. The NEFA model is applied to visual-based space manipulator to trajectory prediction in this measuring instrument. The experimental result shows that the smaller the r has, the higher predictive accuracy the NEFA model has. Besides, the accuracy of 0.1-order NEFA (0.1-NEFA) model is approximately 30 times higher than that both of the traditional first-order grey model and the least-squares estimation (LSE) method. In all, the 0.1-NEFA model is superior to both the traditional first-order gray model and the LSE method.