Low Phase Shift and Least Squares Optimal FIR Filter

Abstract

The problem of a significant phase shift in a control loop is posing a lot of challenges to the control design. One of them is definitely the loss of performance with the increased phase shift when using a filter at the system’s output. This paper contains a description of a new FIR weights determination method focused on low-pass filter design. The primary goal of this method is to minimize the phase shift caused by the filter. The filter theoretically fits a defined polynomial to an asymmetric data set. In this case, the nearest neighbour samples are only taken from the past side of the filtered vector of signal samples. This feature allows reducing the value of the phase shift, especially for a low-frequency spectrum. Therefore, the filter can be used directly in the closed-loop control and will minimize the loss of system performance.