A Review on Tongue Based Assistive Technology, Devices and FPGA Processors Using Machine Learning Module

Prasanna, Dasari Lakshmi; Tripathi, Suman Lata

doi:10.1007/s11277-024-10897-8

Cited by 11 publications

(1 citation statement)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also, EMI filters can integrate with AI technology to accurately perform audio classification in identifying data searches in devices. Several kinds of research were carried out for audio classification, clustering and extraction using machine learning modules [27][28][29][30][31][32] like support vector machine, Naive Bayes, discrete tree, etc.…”

Section: Future Scopementioning

confidence: 99%

High attenuation electromagnetic interface filter for effective processing of audio signals

Prasanna,

Tripathi,

Mahmud

et al. 2024

Eng. Res. Express

View full text Add to dashboard Cite

Suppressing the noise or unwanted signal from the raw signal makes the system work efficiently. Electromagnetic interface (EMI) can cause the system/device to lose the data, damage the electronic equipment, interrupt the audio, or video signals, lead to poor reception, etc. EMI filters are used in telecommunications, military equipment, satellite communication, etc. At the industry level, EMI filters are used for home appliances, medical equipment, motor controls, test equipment, etc. In this paper, active and passive EMI filters are designed to avoid or attenuate EMI or noise (unwanted EMI signal) and implemented on Xilinx Nexys 4 Artix 7 FPGA board. The filters provide a significant improvement in the attenuation of the EMI. MATLAB tool is used to design the filters, which are highly immune to undesired signals, give better stability, and perform over a wide range of frequencies. Depending on the filters' coefficients, the frequency response will be changed. Phase response and amplitude response of the EMI filters are calculated. Insertion loss of the active filter is 1.5836, and for the passive filter, 1.9382. The gain of active and passive filters is 0.7359 and 0.912 respectively. The return loss of the active and passive EMI filters is 20dB. HDL code is generated using MATLAB for the implementation of filter on FPGA. Using MATLAB HDL coder for implementing is new and it simplifies the design, as a result development time is reduced. Xilinx Vivado tool is used for the implementation of active filter. The total on-chip power is 0.146W. Dynamic power is 0.049W, and device static power is 0.097W. The utilization of LUTs and slice registers are less, compared to the previous filter implementation, which leads to reduce the cost of hardware implementation.

show abstract