Effects of Noise Correlation on LEAST SQUARES Filtering in Multipath Detection for GNSS

In GNSS (Global Navigation Satellite System) multipath (MP) results to be one of the main error sources affecting the GNSS solution. In this paper a Linear Adaptive Filter (LAF) technique [1] is applied, based on Least Squares (LS), to estimate the MP coefficients and delay by using a post-correlation approach. An assumption using LAFs [1] is the noise to be a white process, but considering post-correlation data the hypothesis of uncorrelation among the samples is not valid. The LAF is a stat-of-the-art technique, but not in the GNSS-MP-detection and mitigation field. With the objective of using this method for this purpose, the effects of the noise correlation in LS filters are studied in this paper, when the technique is applied to GNSS channel estimate in post-correlation. In this paper, a preliminary analysis is done, by means of simulations. Comparisons are shown between data affected by correlated and uncorrelated noise, using realistic GNSS data