From 378643c8c3174e7ef393061f42a4c598cb2eca5e Mon Sep 17 00:00:00 2001 From: Dhruva Shaw Date: Thu, 12 Dec 2024 19:19:17 +0530 Subject: [PATCH] some changes --- _projects/mcba.md | 58 ++++++++++++++++++++++++++++++++-- assets/bibliography/papers.bib | 56 +++++++++++++++++++++++++++++++- 2 files changed, 111 insertions(+), 3 deletions(-) diff --git a/_projects/mcba.md b/_projects/mcba.md index bcb663f..91c7f16 100644 --- a/_projects/mcba.md +++ b/_projects/mcba.md @@ -4,6 +4,7 @@ title: Mind Controlled Bionic Arm with Sense of Touch description: Imagine a prosthetic arm that functions like your natural arm. You wear a headband, and with the thought process, the working signal from mind connects to the prosthetic about moving the arm, it responds accordingly—just like your real arm! tags: Bionic Arm Robotics Biotechnology Mind Control Prosthetics giscus_comments: true +citation: true img: /assets/img/mcba_logo.jpeg date: 2024-12-12 featured: true @@ -43,7 +44,6 @@ authors: name: Lovely Professional University bibliography: papers.bib -citation: true # Optionally, you can add a table of contents to your post. # NOTES: @@ -54,7 +54,7 @@ citation: true toc: true --- -# Abstract +## Abstract Advancements in bionic technology are transforming the possibilities for restoring hand function in individuals with amputations or paralysis. This paper introduces a cost-effective bionic arm design that leverages mind-controlled functionality and integrates a sense of touch to replicate natural hand movements. The system utilizes a non-invasive EEG-based control mechanism, enabling users to operate the arm using brain signals processed into PWM commands for servo motor control of the bionic arm. Additionally, the design incorporates a touch sensor (tactile feedback) in the gripper, offering sensory feedback to enhance user safety and dexterity. The proposed bionic arm prioritizes three essential features: @@ -62,3 +62,57 @@ The proposed bionic arm prioritizes three essential features: 2. Mind-Control Potential: Harnessing EEG signals for seamless, thought-driven operation. 3. Non-Invasive Nature: Ensuring user comfort by avoiding invasive surgical procedures. This novel approach aims to deliver an intuitive, natural, and efficient solution for restoring complex hand functions. + +--- + +## Methodology +### 1. Data Collection and Dataset Overview +The model development utilized a publicly available EEG dataset comprising data from 60 volunteers performing 8 distinct activities [3]. The dataset includes a total of 8,680 four-second EEG recordings, collected using 16 dry electrodes configured according to the international 10-10 system [3]. +• Electrode Configuration: Monopolar configuration, where each electrode's potential was measured relative to neutral electrodes placed on both earlobes (ground references). +• Signal Sampling: EEG signals were sampled at 125 Hz and preprocessed using: + - A bandpass filter (5–50 Hz) to isolate relevant frequencies [3]. + - A notch filter (60 Hz) to remove powerline interference [3]. + +### 2. Data Preprocessing +The dataset, originally provided in CSV format, underwent a comprehensive preprocessing workflow: +• The data was split into individual CSV files for each of the 16 channels, resulting in an increase from 74,441 files to 1,191,056 files. +• Each individual channel's EEG data was converted into audio signals and saved in .wav format, allowing the brain signals to be audibly analyzed. +• The entire preprocessing workflow was implemented in Python to ensure scalability and accuracy. +The dataset captured brainwave signals corresponding to the following activities: +1) BEO (Baseline with Eyes Open): One-time recording at the beginning of each run [3]. +2) CLH (Closing Left Hand): Five recordings per run [3]. +3) CRH (Closing Right Hand): Five recordings per run [3]. +4) DLF (Dorsal Flexion of Left Foot): Five recordings per run [3]. +5) PLF (Plantar Flexion of Left Foot): Five recordings per run [3]. +6) DRF (Dorsal Flexion of Right Foot): Five recordings per run [3]. +7) PRF (Plantar Flexion of Right Foot): Five recordings per run [3]. +8) Rest: Recorded between each task to capture the resting state [3] [4]. + +### 3. Feature Extraction and Classification +Feature extraction and activity classification were performed using transfer learning with YamNet [5], a deep neural network model. +• Audio Representation: Audio files were imported into MATLAB using an Audio Datastore [6]. Mel-spectrograms, a time-frequency representation of the audio signals, were extracted using the yamnetPreprocess [7] function [8]. +• Dataset Split: The data was divided into training (70%), validation (20%), and testing (10%) sets. +Transfer Learning with YamNet [5] [8]: +- The pre-trained YamNet model (86 layers) was adapted for an 8-class classification task: + -> The initial layers of YamNet [5] were frozen to retain previously learned representations [8]. + -> A new classification layer was added to the model [8]. +- Training details: + -> Learning Rate: Initial rate of 3e-4, with an exponential learning rate decay schedule [8]. + -> Mini-Batch Size: 128 samples per batch. + -> Validation: Performed every 651 iterations. + +### 4. Robotic Arm Design and Simulation +A 3-Degree-of-Freedom (DOF) robotic arm was designed using MATLAB Simulink and Simscape toolboxes. To ensure robust validation: +• A virtual environment was developed in Simulink, simulating the interactions between the trained AI models and the robotic arm. +• The simulations served as a testbed to evaluate the system's performance before real-world integration. + +### 5. Project Progress and Future Directions +Completed Tasks: +1. AI Model Development: Successfully trained models to classify human activities based on EEG signals. +2. Robotic Arm Design: Designed a functional 3-DOF robotic arm with simulated controls. +3. Virtual Simulation: Validated AI-robotic arm interactions in a virtual environment. + +Future Directions: +1. Hardware Integration: Implement the developed AI models into physical robotic hardware for real-world testing. +2. Real-Time EEG Acquisition: Develop a system for real-time EEG data acquisition and activity classification. +3. Tactile Feedback System: Integrate tactile sensors with the robotic arm for real-world sensory feedback, complemented by Simulink-based simulations. diff --git a/assets/bibliography/papers.bib b/assets/bibliography/papers.bib index d8c8c67..7ba3d7e 100644 --- a/assets/bibliography/papers.bib +++ b/assets/bibliography/papers.bib @@ -1,4 +1,58 @@ --- --- +@misc{nprnews, + url={https://www.npr.org/sections/health-shots/2021/05/20/998725924/a-sense-of-touch-boosts-speed-accuracy-of-mind-controlled-robotic-arm}, + journal={NPR}, + year={2021}, + month={May}, + title={Scientists Bring The Sense Of Touch To A Robotic Arm} +} -‌ +‌@misc{transferlearning_matlab, +url={https://in.mathworks.com/help/audio/ug/transfer-learning-with-pretrained-audio-networks.html}, + journal={Mathworks.com}, + year={2024}, + title={Transfer Learning with Pretrained Audio Networks} +} + +‌@misc{classify_sounds_using_yamnet_2021, + url={https://in.mathworks.com/help/audio/ref/yamnetpreprocess.html}, + journal={Mathworks.com}, + year={2021}, + title={yamnetPreprocess} +} + +‌@misc{audio_datastore, + url={https://in.mathworks.com/help/audio/ref/audiodatastore.html}, + journal={Mathworks.com}, + year={2021}, + title={audioDatastore} +} + +‌@misc{yamnet_github, + url={https://github.com/tensorflow/models/tree/master/research/audioset/yamnet}, + journal={GitHub}, + year={2024}, + author={Google and Ellis, Dan and Plakal, Manoj}, +} + +@article{asanza_2023, + title={MILimbEEG: An EEG Signals Dataset based on Upper and Lower Limb Task During the Execution of Motor and Motorimagery Tasks}, volume={2}, url={https://data.mendeley.com/datasets/x8psbz3f6x/2}, + DOI={https://doi.org/10.17632/x8psbz3f6x.2}, + journal={Mendeley Data}, + author={Asanza, Victor and Montoya, Daniel and Lorente-Leyva, Leandro Leonardo and Peluffo-Ordóñez, Diego Hernán and González, Kléber}, + year={2023}, + month={July} +} + +‌@misc{https://doi.org/10.5524/100295, + doi = {10.5524/100295}, + url = {http://gigadb.org/dataset/100295}, + author = {Cho, Hohyun and Ahn, Minkyu and Ahn, Sangtae and {Moonyoung Kwon} and Jun, Sung Chan}, + keywords = {ElectroEncephaloGraphy(EEG), Motor imagery, EEG, brain computer interface, performance variation, subject-to-subject transfer}, + language = {en}, + title = {Supporting data for "EEG datasets for motor imagery brain computer interface"}, + publisher = {GigaScience Database}, + year = {2017}, + copyright = {CC0 1.0 Universal} +}