Wheelchair Free Hands Navigation Using Robust DWT_AR Features Extraction Method With Muscle Brain Signals

Abstract

Researchers try to help disabled people by introducing some innovative applications to support and assess their life. The Brain-Computer Interface (BCI) application that covers both hardware and software models, is considered in this work. BCI is implemented based on brain signals to be converted to commands. To increase the number of commands, non-brain source signals are used, such as eye-blinking, teeth clenching, jaw squeezing, and other movements. This paper introduced a low dimensions robust method to detect the eye-blinks and jaw squeezing; so that the method can be applied to drive a wheelchair by using five commands. Our approach is used Discrete Wavelet Transform with Autoregressive to extract the signal's features. These features are classified by using a linear Support Vector Machine (SVM) classifier. The present method detects every testing sample using a small training set to test and drive a powered wheelchair. The proposed method is fully implemented practically based on binary-coded commands.

Deep-Learning Approach for Efficient Eye-blink Detection with Hybrid Optimization Concept

“…Pr 1 (32) The MSE of the proposed work is the lowest one for every variation in the learning percentage. The MSE of the projected work is found be below 8% for every variation in the learning percentage.…”

“…The proposed work has been implemented in PYTHON. The dataset for evaluation has been collected from [32]. The collected sample images are shown in Fig.…”

“…In addition, the projected model with optimization has recorded the least MSLE and RMSE also. Thus, from the overall evaluation, it's obvious that the major improvement behind the projected work in precisely detecting the eye blinks is only due to the fine-tuning of the weight of CNN via the optimization model [32].…”

A Comprehensive Review of Endogenous EEG-Based BCIs for Dynamic Device Control

“…The majority of the papers we reviewed obtained the control signal directly from the output of the BCI classifier [ 2 , 28 , 31 , 33 , 37 , 43 , 46 , 48 , 50 , 56 , 59 , 60 , 61 , 63 , 66 , 72 ]. For asynchronous systems, this generally involved windowing the EEG data and obtaining a classification output at a regular rate [ 2 , 28 , 33 , 37 , 43 , 46 , 48 , 50 , 56 , 59 , 60 , 61 , 63 , 66 , 71 , 72 ]. This classification output was used to drive the external device.…”

“…Many studies used CSP features to capture spatiotemporal characteristics in the data, and then paired them with time-domain features [ 37 , 45 ], time-frequency domain features such as the discrete wavelet transform [ 40 , 84 ], and/or functional brain network features [ 54 ]. Other studies paired time-domain features with time-frequency domain features [ 60 , 72 , 85 ] or frequency domain features [ 61 ].…”