Research on coding and decoding algorithm of binocular brain-controlled unmanned vehicle

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Steady-state visual evoked potential (SSVEP) is effective for brain-computer interfaces (BCIs).
• Traditional single-frequency SSVEP has limitations in command scalability and visual comfort.
• Brain-controlled unmanned vehicles (UVs) require efficient and comfortable BCI solutions.

Purpose of the Study:

• To develop a novel binocular SSVEP stimulation paradigm for enhanced UV control.
• To improve command set scalability and visual comfort in SSVEP-based BCIs.
• To introduce and validate an improved filter bank dual-frequency task-discriminant component analysis (FBD-TDCA) algorithm.

Main Methods:

• A binocular SSVEP stimulation paradigm using checkerboard and phase encoding with dual frequencies per target (30-35 Hz).
• Polarized light technology to present distinct frequencies to each eye, reducing visual interference.
• An improved filter bank dual-frequency task-discriminant component analysis (FBD-TDCA) algorithm for signal processing.

Main Results:

• Six frequencies encoded 15 commands with performance comparable to traditional methods.
• The FBD-TDCA algorithm achieved 89.27% ± 3.67 classification accuracy and 163.87 ± 14.32 bits min⁻¹ information transfer rate.
• Online 12-command UV control task showed 90.34% ± 8.75% accuracy with low path deviation.

Conclusions:

• The proposed binocular SSVEP paradigm enhances command scalability and visual comfort.
• The FBD-TDCA algorithm offers superior performance over existing methods.
• This approach advances efficient and user-friendly BCI applications for real-world scenarios like UV control.