e-space
Manchester Metropolitan University's Research Repository

Regulated-element Frost Beamformer for Vehicular Multimedia Sound Enhancement and Noise Reduction Applications

Ekpo, Sunday and Adebisi, B and Wells, A (2017) Regulated-element Frost Beamformer for Vehicular Multimedia Sound Enhancement and Noise Reduction Applications. IEEE Access, 5. pp. 27254-27262. ISSN 2169-3536

[img]
Preview

Available under License ["licenses_description_i" not defined].

Download (7MB) | Preview

Abstract

A key requirement of an adaptive sensor array involves the ability to deterministically adjust the directional response of the array to reduce noise and reverberations, null interferences and enhance the gain and recognition of the desired signal. This paper presents a low-carbon adaptive broadband beamforming algorithm called the regulated-element Frost beamformer. It enhances the desired signal based on the noise conditions of the individual omnidirectional sensors deployed in a complex dynamic environment that is prone to steering errors. The investigation of this algorithm was carried out in an interference-dominant, noisy automobile environment characterised by diffuse noise conditions. An embedded system measurement of real-time signals was carried out using omnidirectional acoustic sensors mounted in a model convertible F-Type car driven at speed limits of 20 to 50 mph. The simulation results indicate an array gain enhancement of 2 dB higher than the conventional Frost beamformer and it requires less sensors and filter taps for real-time reconfigurable implementations. The experimental results reveal that the average array gain of the regulated-element beamformer is 2.9 dB higher than the conventional Frost beamformer response. The minimum floor array gain of the regulated-element beamformer is 5 dB, representing 70 % noise reduction than the conventional adaptive beamformers.

Impact and Reach

Statistics

Downloads
Activity Overview
24Downloads
51Hits

Additional statistics for this dataset are available via IRStats2.

Altmetric

Actions (login required)

Edit Item Edit Item