2019Unpublished venueRequires access

Mismatched Filtering of Doppler Tolerant Codes for Multi-code Sonar Systems

Muhammad Salman Rashed, Marco Meijer, Paul D. Teal

Open publisher page 1 citations

Abstract

Sets of transmitted signals used for sonar sensing applications must have good correlation properties, conformance of the magnitude to a target profile, and Doppler tolerance. Achieving an optimal solution with all three constraints along with the receiver, transmitter and channel effects is a complex and non-convex problem. Whereas optimising signals for the first two constraints is relatively easy, a Doppler shift degrades the sidelobe levels considerably so that achieving low and flat sidelobe levels is not guaranteed with the use of a matched filter. An alternative is to design a mismatched filter for reducing the sidelobe levels beyond the limits achievable with a matched filter. Designing a mismatched filter which is longer in length than a matched filter offers greater flexibility in terms of the filter coefficients that can be tuned. Designing a signal may be a non-convex problem, but with the signal fixed, design of a mismatched filter is convex. We compare two different methods for designing mismatched filters for Doppler tolerant codes. Results from both our designs show flat sidelobes at least 2 dB lower than those of a matched filter.

About this research paper

What this paper is about

Sets of transmitted signals used for sonar sensing applications must have good correlation properties, conformance of the magnitude to a target profile, and Doppler tolerance. Achieving an optimal solution with all three constraints along with the receiver, transmitter and channel effects is a complex and non-convex problem. Whereas optimising signals for the first two constraints is relatively easy, a Doppler shift degrades the sidelobe levels considerably so that achieving low and flat sidelobe levels is not guaranteed with the use of a matched filter. An alternative is to design a mismatched filter for reducing the sidelobe levels beyond the limits achievable with a matched filter. Designing a mismatched filter which is longer in length than a matched filter offers greater flexibility in terms of the filter coefficients that can be tuned. Designing a signal may be a non-convex problem, but with the signal fixed, design of a mismatched filter is convex. We compare two different methods for designing mismatched filters for Doppler tolerant codes. Results from both our designs show flat sidelobes at least 2 dB lower than those of a matched filter.

Why it matters

OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Sets of transmitted signals used for sonar sensing applications must have good correlation properties, conformance of the magnitude to a target profile, and Doppler tolerance. Achieving an optimal solution with all three constraints along with the receiver, transmitter and channel effects is a complex and non-convex problem. Whereas optimising signals for the first two constraints is relatively easy, a Doppler shift degrades the sidelobe levels considerably so that achieving low and flat sidelobe levels is not guaranteed with the use of a matched filter. An alternative is to design a mismatched filter for reducing the sidelobe levels beyond the limits achievable with a matched filter. Designing a mismatched filter which is longer in length than a matched filter offers greater flexibility in terms of the filter coefficients that can be tuned. Designing a signal may be a non-convex problem, but with the signal fixed, design of a mismatched filter is convex. We compare two different methods for designing mismatched filters for Doppler tolerant codes. Results from both our designs show flat sidelobes at least 2 dB lower than those of a matched filter.

Key concepts: Matched filter, Filter (signal processing), Root-raised-cosine filter, Filter design, Sonar, Doppler effect, Transmitter, Computer science

Related papers

Back to paper searchBrowse research topicsOriginal source
Mismatched Filtering of Doppler Tolerant Codes for Multi-code Sonar Systems — Research Paper | ScholarLens