Enhanced Digital Acoustic Perception in Hearing Aid Device Using Reconfigurable Filter Bank Structure: A Systematic Review with Recommendation

Rekha  Karuppaiah; Umadevi  Seerengasamy; Nagajayanthi  Boobalakrishnan

doi:10.18502/avr.v34i2.18052

Rekha Karuppaiah School of Electronics Engineering, Vellore Institute of Technology, Chennai Campus, Chennai, India
Umadevi Seerengasamy Centre for Nanoelectronics and VLSI Design, Vellore Institute of Technology, Chennai Campus, Chennai, India
Nagajayanthi Boobalakrishnan School of Electronics Engineering, Vellore Institute of Technology, Chennai Campus, Chennai, India

DOI: https://doi.org/10.18502/avr.v34i2.18052

Abstract

Background and Aim: Untreated hearing loss can severely impact quality of life, mental and physical health, and cognitive performance. Digital hearing aids can mitigate these effects, with the filter bank being a crucial component. It divides signals into frequency bands, compresses, amplifies, and processes speech based on the user’s hearing profile. This study focused on optimizing filter bank architecture in terms of hardware cost, processing speed, and adaptability to enhance the efficiency of digital hearing aids.

Recent Findings: Each filter bank in digital hearing aids relies on Finite Impulse Respons (FIR) filters, and optimizing their architecture is crucial for optimal device performance. Literature suggests that reconfigurable digital FIR filters are preferred for filter bank structures. However, their performance may vary based on specifications such as filter length, bandwidth, sampling frequency, and coefficients. Therefore, this review aimed to identify an optimized reconfigurable FIR filter design that improves hearing aid performance while ensuring its parameters remain independent of these specifications.

Conclusion: A hardware-efficient, optimized, and adaptable parallel computing architecture for hearing aid filter banks has been identified from the literature survey. This proposed architecture features reconfigurable sub-band frequencies tailored to the user’s specific hearing loss, utilizing a Coefficient Scanning Mechanism (CSM) and Floating Point-Computation Sharing High-speed Mechanism (FP-CSHM). The CSM dynamically adjusts sub-band selection and reorganizes the FIR structure in each filter bank to reduce multiplication counts based on coefficient matching. The FP-CSHM enhances computation speed by eliminating redundant calculations through parallel processing.

Keywords: Hearing aid; hearing loss; audiogram; hearing threshold