There could be large error in Compressive Sensing Ultra-Wide Band (CS-UWB) channel estimation because of the coherence of sensing information matrix and the noise interference.Firstly

the Adaptive Sensing Information (ASI) matrix is constructed by measurement vector weighting

which has the desirable properties of low coherence and containing measurement information to support the sparse recovery algorithms to obtain optimal atoms.Secondly

a modified Sparsity Adaptive Matching Pursuit (SAMP) algorithm is proposed to recovery sparse signal accurately without prior information of sparsity or SNR.Finally

a non-convex CS-UWB channel estimation method is presented based on the ASI matrix and the modified SAMP algorithm.Both the theoretical analysis and the experimental results show that this method can accurately reconstruct UWB communication channel in low SNR conditions with very low compressive sampling ratio.