Analog filter adaptation using a dithered linear search algorithm

A variation of the differential steepest descent algorithm, here called the dithered linear search (DLS), is examined and applied to analog filter adaptation. The DLS algorithm is a gradient descent optimizer with a straightforward and robust hardware implementation. Gradient estimates are obtained by applying independent additive dither to all of the filter's parameters simultaneously and correlating the resulting changes in the output squared error to the dither signals. Unlike the popular LMS algorithm, the DLS algorithm does not require access to the filter's internal states. No additional analog hardware is required making it ideal for adaptive analog filters in mixed-signal systems. A theoretical analysis shows no gradient misalignment. The algorithm is verified on an integrated analog filter. The effects of dc offsets are also examined.