Unstable CIR Values

Im currently using two dw1000 and plot out their phases at their respective first path index to try to estimate the pdoa of the same received frame, but both first path phases jump around even in a static environment. Is there any way to stabilize those readings?