Technical Projects and ResearchÂ
Highlights in innovation and contributions to technical advancements in healthcare
Highlights in innovation and contributions to technical advancements in healthcare
Addressing the Core Challenge in Bionic Device Adoption
While modern bionic devices offer remarkable technical capabilities, nearly 50% of patients abandon them within months. Research indicates that patient-tailored neurotraining is critical for long-term adherence. This project tackled this issue by automating and enhancing the personalization of neurotraining exercises through an interactive game-based interface.
A validated real-time EMG-controlled game interface that adjusts difficulty dynamically based on user performance.
🔑Key Features:
Calibration Stage: Acquisition and analysis of myoelectric performance indicators.
Automated Personalization: Real-time optimization using cosine similarity algorithms.
Interactive Interface: Immersive game experience controlled via EMG signals.
Progress Assessment: Potential of data-driven tracking of user improvement over time.
🏆 Recognition: Presentation of the project to be published at the IEEE Cyborg and Bionic Systems Conference 2024.
🛠️ Tools & Technologies:
Hardware: EMG Sensing Techology by Myoware, Teardrop Gel Electrodes.
Programming Languages: Python, C++.
Techniques: Signal Processing, Real-Time Adaptation, Cosine Similarity Algorithms
Accessibility: Reduces reliance on specialized facilitators.
Scalability: Enables bionic companies to expand prosthetic placements without compromising training quality.
Human-Centered Design: Automates repetitive drills, allowing facilitators to focus on patient connection and care.
Project sponsored by Mirai Innovation Research Institute and supported by Kawatek Co., Ltd. Osaka, Japan, 2023-2024.
Real-Time Adaptation in Immersive Virtual Therapies
Psychological therapies in virtual reality hold immense potential; however, their effectiveness hinges on user immersion, which is closely tied to real-time adaptation to their emotional state. Galvanic Skin Response (GSR) has proven to be a reliable biosignal to detect fluctuations in a subject's state. Therefore, this set of projects leveraged GSR analysis to dynamically trigger interactive elements in virtual environments, with the aim to enhance therapeutic outcomes in areas such as phobia treatment and relaxation techniques.Â
A robust real-time GSR analysis system capable of adapting virtual environments dynamically based on user stress levels.
🔑Key Contributions:
Biosignal Acquisition Protocol: Optimized sampling protocol with automated documentation for efficient analysis.
Feature Extraction: Scripts for data processing and real-time extraction of stress indicators.
Adaptive Environment Triggers: Real-time feedback loop to activate stress-specific interactive elements in virtual scenarios.
Scalable Design: Framework extended to multiple rehabilitation projects within Mirai Innovation Research Institute.
🛠️ Tools & Technologies:
Hardware: GSR Biosignal Sensing by Groove and Quest 3 Headset by Meta.
Programming Languages: Python, MATLAB.
Techniques: Signal Processing, Feature Extraction, Real-Time Adaptation.
🌍 Project Impact:
Quantifiable Results: Bridging qualitative therapy assessments with quantitative biosignal data.
Therapeutic Personalization: Enabling dynamic adjustments for improved psychological intervention outcomes.
Scalability: Expanded adoption across multiple projects, reinforcing GSR’s role in healthcare-focused immersive technologies.
Project elaborated for Mirai Innovation Research Institute. Osaka, Japan, 2024.
Optimized EEG Sampling Protocol for Real-Time Relaxation DetectionÂ
As mental well-being becomes a priority in health initiatives worldwide, real-time detection of relaxation through EEG signals presents a unique opportunity to enhance virtual health environments. In collaboration with Mirai Innovation Research Institute, this project automated EEG data acquisition and processing to create an adaptive virtual environment for real-time relaxation detection, set to be showcased at the Health Pavilion during Expo Osaka 2025.Â
A state-of-the-art EEG signal processing protocol capable of detecting relaxation stages with real-time feedback, optimized for immersive virtual environments.
🔑Key Features:
Automated EEG Data Acquisition: Streamlined process for continuous data sampling across multiple stages of cognitive and emotional engagement.
Signal Processing: Advanced filtering, segmentation, and feature extraction techniques for accurate real-time analysis.
Virtual Environment Integration: Synchronization of EEG sampling with virtual environments to assess relaxation and engagement.
Relaxation Detection: Utilization of Relative Gamma Power (RG) to characterize relaxation and cognitive states.
🛠️ Tools & Technologies:
Hardware: EEG Technology (AURA, Mirai Innovation Research Institute) with LabStream Layer for data acquisition.
Programming Languages: Python, MATLAB.
Techniques: Signal Filtering (Butterworth IIR, Notch Filter), Epoch Segmentation, Spectral Analysis (Welch’s Method), Feature Extraction.
🌍 Project Impact:
Health Innovation: Offers new ways to monitor and enhance relaxation in real-time, benefiting health and wellness programs.
Scalability: Can be adapted for use in various health and rehabilitation settings globally, especially in interactive environments like exhibitions.
Data-Driven Insights: Facilitates the tracking of brain activity during stress, engagement, and relaxation.
Project elaborated for Mirai Innovation Research Institute. Osaka, Japan, 2024.
Enhancing Prosthetic Usability through AI-Driven EMG Signal Classification
EMG-based gesture recognition has emerged as a transformative approach, enabling users to interact with prosthetics in a more natural and seamless way. This project, originating from a self-developed workshop, showcased the power of AI in bionic systems, demonstrating how real-time hand gesture predictions—relaxation, opening, and closing—can be achieved using a single electrode and advanced AI modeling.Â
An AI-powered EMG system enabling real-time hand gesture recognition for enhanced prosthetic control.Â
🔑Key Features:Â
Optimized Dataset Creation: Robust labeled dataset tailored for accurate hand gesture prediction.
Real-Time Gesture Detection: Precise classification of relaxation, opening, and closing states.
Streamlined AI Integration: Adaptation of pre-existing scripts for efficient real-time predictions.
🛠️ Tools & Technologies:
Libraries: Python (NumPy, scikit-learn, PyTorch).
Techniques: Real-Time Signal Processing, State Classification, Data Labeling.
🌍 Project Impact:
User-Centric Design: Creates a more natural connection between the user and their prosthetic device, transforming it into a seamless extension of their body.
Scalable Integration: Foundational framework for future AI-based prosthetic systems.
Project elaborated for Mirai Innovation Research Institute. Osaka, Japan, 2024.