Dr. Hamed Ghane

Project Overview

This research developed a novel control algorithm based on a P300-based brain-computer interface (BCI) to control a 2-DoF robotic arm. The system allows users to control robotic movement through brain signals alone, providing a potential pathway for assistive technology for patients with motor impairments such as Amyotrophic Lateral Sclerosis (ALS) and Spinal Cord Injury (SCI).

Eight subjects (five men and three women) performed a 2-dimensional target tracking task in a simulated environment. Their EEG signals from the visual cortex were recorded, and P300 components were extracted and evaluated to deliver a real-time BCI-based controller. The volunteer's intention was recognized and decoded as an appropriate command to control the cursor, which was then used to control a simulated robotic arm in a 2-dimensional space.

Methodology

System Architecture

The BCI system was implemented in four key steps:

1. Visual Stimulus Design: A custom GUI with four bulbs representing directions (up, down, left, right) that flash randomly to elicit P300 responses
2. EEG Signal Processing: Recording brain signals from the visual cortex followed by filtering and feature extraction
3. Classification: Using K-nearest neighbors (KNN) and support vector machine (SVM) classifiers to interpret brain signals
4. Robotic Control: Converting classified brain signals into movement commands for the robotic arm

Feature Extraction and Classification

The signal processing pipeline included:

• Hilbert-Huang Transform: For non-linear and non-stationary signal analysis
• Wavelet Transform: For time-frequency analysis of EEG signals
• Principal Component Analysis (PCA): For dimension reduction
• K-nearest Neighbors (KNN) Classification: Achieving a 97% recognition rate

Robotic Control System

The robotic arm was controlled using an inverse dynamics control method:

• Dynamics Modeling: Accurate modeling of the 2-DoF robotic arm dynamics
• Inverse Dynamics Control: Converting desired trajectories into appropriate joint torques
• Reference Input Generation: Translating BCI outputs into reference positions for the robot
• Point-to-Point Control: Enabling smooth movement between arbitrary positions

Results & Applications

Drawing Task Demonstration

To validate the system's practical application, subjects were asked to draw shapes using the BCI-controlled robotic arm:

1. The subject imagines the desired shape and the required movements
2. For each movement, the subject focuses on the corresponding direction's bulb in the BCI panel
3. The BCI system detects the P300 response and translates it into a direction command
4. The robotic arm moves the end-effector (with attached marker) accordingly
5. By sequentially focusing on different directions, complete shapes can be drawn

Impact and Applications

This research demonstrates the potential for BCI systems to provide intuitive control mechanisms for robotic devices, particularly for:

Interested in Collaboration?

This research demonstrates the potential of BCI technology for robotic control applications. If you're interested in brain-computer interfaces or need a neural control system for assistive technology, let's discuss how we can work together.