Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Driver-Intention Prediction with Deep Learning: Real-Time Brain-to-Vehicle Communication

Niloufar Alavi, Swati Shah, Rezvan Alamian, Stefan Goetz

TL;DR

This work tackles real-time driver intention prediction by decoding steering intent from noninvasive EEG signals. It introduces a 1D CNN that operates on raw EEG with minimal pre-processing, trained on a driving-simulator dataset to classify left, straight, and right steering intentions with $83.7\%$ accuracy. The study reveals a notable spatial bias (best performance for right-turn imagery) and identifies timing windows and spectral patterns (notably in the alpha and beta bands) that differentiate conditions, while acknowledging potential gaze confounds. Collectively, the results demonstrate the feasibility of EEG-based brain-to-vehicle communication for driving assistance and highlight avenues for improving robustness through larger, session-spanning datasets and addressing subject-specific factors.

Abstract

Brain-computer interfaces (BCIs) allow direct communication between the brain and electronics without the need for speech or physical movement. Such interfaces can be particularly beneficial in applications requiring rapid response times, such as driving, where a vehicle's advanced driving assistance systems could benefit from immediate understanding of a driver's intentions. This study presents a novel method for predicting a driver's intention to steer using electroencephalography (EEG) signals through deep learning. A driving simulator created a controlled environment in which participants imagined controlling a vehicle during various driving scenarios, including left and right turns, as well as straight driving. A convolutional neural network (CNN) classified the detected EEG data with minimal pre-processing. Our model achieved an accuracy of 83.7% in distinguishing between the three steering intentions and demonstrated the ability of CNNs to process raw EEG data effectively. The classification accuracy was highest for right-turn segments, which suggests a potential spatial bias in brain activity. This study lays the foundation for more intuitive brain-to-vehicle communication systems.

Driver-Intention Prediction with Deep Learning: Real-Time Brain-to-Vehicle Communication

TL;DR

This work tackles real-time driver intention prediction by decoding steering intent from noninvasive EEG signals. It introduces a 1D CNN that operates on raw EEG with minimal pre-processing, trained on a driving-simulator dataset to classify left, straight, and right steering intentions with accuracy. The study reveals a notable spatial bias (best performance for right-turn imagery) and identifies timing windows and spectral patterns (notably in the alpha and beta bands) that differentiate conditions, while acknowledging potential gaze confounds. Collectively, the results demonstrate the feasibility of EEG-based brain-to-vehicle communication for driving assistance and highlight avenues for improving robustness through larger, session-spanning datasets and addressing subject-specific factors.

Abstract

Brain-computer interfaces (BCIs) allow direct communication between the brain and electronics without the need for speech or physical movement. Such interfaces can be particularly beneficial in applications requiring rapid response times, such as driving, where a vehicle's advanced driving assistance systems could benefit from immediate understanding of a driver's intentions. This study presents a novel method for predicting a driver's intention to steer using electroencephalography (EEG) signals through deep learning. A driving simulator created a controlled environment in which participants imagined controlling a vehicle during various driving scenarios, including left and right turns, as well as straight driving. A convolutional neural network (CNN) classified the detected EEG data with minimal pre-processing. Our model achieved an accuracy of 83.7% in distinguishing between the three steering intentions and demonstrated the ability of CNNs to process raw EEG data effectively. The classification accuracy was highest for right-turn segments, which suggests a potential spatial bias in brain activity. This study lays the foundation for more intuitive brain-to-vehicle communication systems.
Paper Structure (13 sections, 7 figures, 2 tables)

This paper contains 13 sections, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Research Design
  4. Participants
  5. Procedure
  6. Analysis
  7. Data Training
  8. Visualization
  9. Results
  10. Discussion
  11. Model Size and Variability
  12. Sensitive Intervals and Spectral Characteristics
  13. Spatial Bias

Figures (7)

  • Figure 1: Experimental setup of the driving simulator.
  • Figure 2: Driver's view for (a) straight driving and (b) turning.
  • Figure 3: Configuration of the signal flow in the experimental setup.
  • Figure 4: Accuracy and loss for the generic model.
  • Figure 5: Topographic plots of average brain activity within the first 3 s after the trigger occurrence for (a) left, (b) straight, and (c) right.
  • ...and 2 more figures