Computationally efficient mismatched filter design for pulse compression codes
MetadataShow full item record
In this dissertation, computationally efficient mismatched filters are proposed to achieve sidelobe suppression in pulse compression codes. Length-optimal mismatched filters, that achieve the best possible peak sidelobe suppression or lowest integrated sidelobe levels for a given filter length, have been proposed in the literature. Since all the coefficients in these filters are optimized, a filter of length N , in general, requires N multipliers and N - 1 adders for its implementation. In this work, we propose two different types of mismatched filters that use significantly fewer multipliers and adders than the length-optimal filters for similar levels of sidelobe suppression. This is achieved at the cost of an increased latency compared to the length-optimal filters. As shown in this work, the proposed filters are particularly suitable for pulse compression codes that exhibit some regularity in the structure of their sidelobes. Results are presented for Barker codes and Huffman sequences and extended for compound Barker codes.