Largest Lyapunov Exponent Optimization for Control of a Bionic-Hand: A Brain Computer Interface Study

Area of Science:

• Neuroscience and Biomedical Engineering
• Non-linear Dynamics and Chaos Theory

Background:

• Brain-computer interfaces (BCIs) are crucial for advanced prosthetics.
• Electroencephalography (EEG) signals offer a non-invasive window into brain activity.
• Quantifying brain signal complexity is key for effective BCI control.

Purpose of the Study:

• To enhance the accuracy of brain-controlled bionic hands.
• To optimize the Largest Lyapunov Exponent (LLE) algorithm for EEG signal analysis.
• To compare the performance of Water Drop (WD) and Chaotic Tug of War (CTW) optimizers for LLE parameter tuning.

Main Methods:

• Non-invasive EEG recordings were used to capture brain signals.
• The Largest Lyapunov Exponent (LLE) algorithm was employed to quantify brain signal complexity.
• LLE input parameters (false nearest neighbor, mutual information) were optimized using WD and CTW algorithms.
• Optimized LLE was used to decode imaginary movement patterns for bionic hand control.

Main Results:

• The Chaotic Tug of War (CTW) optimizer achieved a superior average accuracy of 72.31% for bionic hand control.
• The CTW method demonstrated higher precision compared to the traditional LLE and the WD-optimized LLE.
• The study confirmed the necessity of optimizing traditional LLE algorithms for improved performance.

Conclusions:

• The optimized LLE, particularly using the CTW method, significantly improves the accuracy of EEG-based bionic hand control.
• The CTW optimizer presents a more efficient and effective approach for LLE parameterization in BCI applications.
• This research highlights the potential of advanced computational methods for developing sophisticated neuroprosthetics.