The prediction issue of image tracker delay and two related improvements

To reduce the impact of the tracker delay in the image tracking system, an equivalent model of the correction method with prepensely delayed turntable angle is reached by mathematic deduction, which shows this method can theoretically eliminate the distortion impact to the tracking response; The remaining impact is just a pure delay to the target angle, which can be described as a special signal prediction issue. And then an improved polynomial fitting and prediction algorithm is proposed. It raises the prediction frequency from the image frame frequency to the turntable angle sensor´s frequency without too much addition to the real-time calculation. Depending on the statistical regularity of polynomial fitting, a further method is proposed using the last residua prediction error. The simulation results show that, the two improvements can be more approximate to the true values and reduce the impact of the tracker delay effectually, and can be applied with other prediction algorithm easily.