Nonholonomic mobile system control by combining EEG-based BCI with ANFIS.

Area of Science:

• Neuroscience and Robotics
• Brain-Computer Interface (BCI) technology
• Human-Computer Interaction

Background:

• Motor imagery electroencephalography (EEG)-based Brain-Computer Interfaces (BCIs) offer non-invasive control for peripheral devices like wheelchairs.
• Current limitations include intensive training requirements for accurate control of complex nonholonomic mobile systems.
• Challenges exist in achieving precise and effective control without user fatigue or external stimulation.

Purpose of the Study:

• To propose a novel approach combining motor imagery EEG with the Adaptive Neural Fuzzy Inference System (ANFIS) for improved mobile system control.
• To enhance the human intelligence fusion with the precision of ANFIS for nonholonomic mobile systems.
• To achieve multi-level control enabling high maneuverability without system halts or reliance on sensor data.

Main Methods:

• Integration of motor imagery EEG signals with an Adaptive Neural Fuzzy Inference System (ANFIS) controller.
• Development of a multi-level control strategy for nonholonomic mobile systems.
• Online training capability of the ANFIS controller during the control task.

Main Results:

• Demonstrated effective multi-level control of a nonholonomic mobile robot.
• Achieved high controllability without requiring the mobile system to stop.
• Verified increased control accuracy and efficiency for the user through ANFIS online training.

Conclusions:

• The proposed approach effectively fuses human motor imagery intent with ANFIS precision for advanced mobile system control.
• This BCI method enhances user experience by enabling intuitive and efficient operation of nonholonomic robots.
• The system's ability to train online significantly boosts control performance and user adaptability.