A Brain-Controlled Vehicle System Based on Steady State Visual Evoked Potentials

Zhang, Zhao; Han, Shuning; Yi, Huaihai; Duan, Feng; Kang, Fei; Sun, Zhe; Solé-Casals, Jordi; Caiafa, Cesar F. · 2022 · Crossref

DOI: 10.1007/s12559-022-10051-1

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the development of a safe, human-vehicle cooperative driving system that integrates a brain-controlled vehicle (BCV) mode with intelligent obstacle detection. The research is motivated by the potential of BCIs to assist drivers with physical disabilities and enhance driving experiences, while acknowledging the safety risks associated with relying solely on BCI commands for outdoor vehicle operation. The authors propose a system where driver intentions are decoded via EEG signals, but vehicle movement is constrained by an independent laser-based obstacle detection module to prevent collisions. The system architecture comprises five main components: an SSVEP-based BCI, a laser ranging obstacle detection system, a computer processing terminal, a communication system, and an intelligent vehicle. The BCI utilizes steady-state visual evoked potentials (SSVEP) elicited by two flickering visual stimuli (8 Hz for braking, 10 Hz for moving) displayed on a screen. EEG signals were acquired non-invasively using a 16-channel g.USBamp amplifier, with electrodes placed at occipital and parietal sites (Oz, O1, O2, POz, PO3, PO4). Signal preprocessing involved 50-Hz notch filtering and Butterworth band-pass filtering (5–60 Hz). Classification was performed using Canonical Correlation Analysis (CCA) combined with an Overlap Time Windows Voting (OTWV) method to improve accuracy without requiring subject-specific training. The obstacle detection system uses a laser sensor to monitor distance; if an obstacle is too close, the system overrides moving commands and triggers an electronic brake switch. Experiments were conducted with five healthy subjects in both simulation and real outdoor vehicle environments. The real vehicle moved at a constant speed of 1.38 m/s when a valid moving command was received and no obstacle was detected. The study verified the feasibility of the system through these trials. The outdoor experimental results demonstrated that the average accuracy of intention recognition on the real vehicle platform was 90.68% ± 2.96%. The system successfully integrated BCI control with safety-critical obstacle avoidance, ensuring that the vehicle would brake if the laser sensor detected proximity to an obstacle, regardless of the BCI command. The significance of this work lies in its demonstration of a feasible, safe human-vehicle cooperative driving system that combines BCI technology with intelligent driving assistance. By integrating obstacle detection, the system mitigates the safety risks inherent in pure BCI-controlled vehicles, particularly in unstructured outdoor environments. The high recognition accuracy and the successful implementation of a training-free classification method suggest that SSVEP-based BCIs can be effectively applied to real-world vehicle control scenarios, providing a supplementary driving mode that could benefit individuals with motor impairments.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• drowsiness detection algorithms

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data
• Methodological Resource: tool software
• Theoretical Contribution: computational model