EEG-based decoding of error-related brain activity in a real-world driving task

Zhang, Huaijian; Chavarriaga, Ricardo; Khaliliardali, Zahra; Gheorghe, Lucian; Iturrate, Iñaki; Millán, José del R.

doi:10.1088/1741-2560/12/6/066028

Cited by 102 publications

(74 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ErrPs have been used in brain-computer interfaces and HRI tasks for performing binary classification, correcting classification errors, and controlling various robot platforms [46,49,41,22,11,29,21,37,38,53,45]. Nevertheless, most ErrP studies are performed in controlled scenarios because EEG signals have low signalto-noise ratios that make real-time classification challenging [40,6,7,31,34,48].…”

Section: A Eeg-based Methods For Human-robot Interactionmentioning

confidence: 99%

Plug-and-Play Supervisory Control Using Muscle and Brain Signals for Real-Time Gesture and Error Detection

DelPreto

Salazar-Gomez

Gil

et al. 2018

Robotics: Science and Systems XIV

View full text Add to dashboard Cite

Abstract-Control of robots in safety-critical tasks and situations where costly errors may occur is paramount for realizing the vision of pervasive human-robot collaborations. For these cases, the ability to use human cognition in the loop can be key for recuperating safe robot operation. This paper combines two streams of human biosignals, electrical muscle and brain activity via EMG and EEG, respectively, to achieve fast and accurate human intervention in a supervisory control task. In particular, this paper presents an end-to-end system for continuous rollingwindow classification of gestures that allows the human to actively correct the robot on demand, discrete classification of Error-Related Potential signals (unconsciously produced by the human supervisor's brain when observing a robot error), and a framework that integrates these two classification streams for fast and effective human intervention. The system also allows "plug-and-play" operation, demonstrating accurate performance even with new users whose biosignals have not been used for training the classifiers. The resulting hybrid control system for safety-critical situations is evaluated with 7 untrained human subjects in a supervisory control scenario where an autonomous robot performs a multi-target selection task.

show abstract

Section: A Eeg-based Methods For Human-robot Interactionmentioning

confidence: 99%

Plug-and-Play Supervisory Control Using Muscle and Brain Signals for Real-Time Gesture and Error Detection

DelPreto

Salazar-Gomez

Gil

et al. 2018

Robotics: Science and Systems XIV

View full text Add to dashboard Cite

show abstract

“…3,4 . Along this line, the possibility of decoding brain activity, measured with electroencephalography (EEG), to improve driving assistance systems has recently attracted considerable interest 11,13,14,29 . In this approach, the objective is not to 'control a car with a BCI' but rather to predict the driver's intentions 11,13,14,29 .…”

Section: Introductionmentioning

confidence: 99%

Real-time Detection of Driver’s Movement Intention in Response to Traffic Lights

Khaliliardali

Chavarriaga

Zhang

et al. 2019

Preprint

View full text Add to dashboard Cite

Movements are preceded by certain brain states that can be captured through various neuroimaging techniques. Brain-Computer Interfaces can be designed to detect the movement intention brain state during driving, which could be beneficial in improving the interaction between a smart car and its driver, by providing assistance in-line with the driver's intention. In this paper, we present an Electroencephalogram based decoder of such brain states preceding movements performed in response to traffic lights in two experiments: in a car simulator and a real car. The results of both experiments (N=10: car simulator, N=8: real car) confirm the presence of anticipatory Slow Cortical Potentials in response to traffic lights for accelerating and braking actions. Single-trial classification performance exhibits an Area Under the Curve (AUC) of 0.71±0.14 for accelerating and 0.75±0.13 for braking. The AUC for the real car experiment are 0.63±0.07 and 0.64±0.13 for accelerating and braking respectively. Moreover, we evaluated the performance of real-time decoding of the intention to brake during online experiments only in the car simulator, yielding an average accuracy of 0.64±0.1. This paper confirm the existence of the anticipatory slow cortical potentials and the feasibility of single-trial detection these potentials in driving scenarios.

show abstract

“…The feasibility of performing single-trial recognition of ErrPs has been demonstrated in different paradigms and setups, including P300 spellers [7], Motor-imagery BCI [6], human-robot interaction [9], [8], and car driving [10]. Noticeably these signals, linked to cognitive monitoring processes, have also been reported to be rather stable across different recording days [11], and feedback characteristics [12], [9].…”

Section: Introductionmentioning

confidence: 99%

Spatial filters yield stable features for error-related potentials across conditions

Iwane

Chavarriaga

Iturrate

et al. 2016

2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC)

View full text Add to dashboard Cite

Abstract-Error-related potentials (ErrP) have been increasingly studied in psychophysical experiments as well as for brainmachine interfacing. In the latter case, the generalisation capabilities of ErrP decoders is a crucial element to avoid frequent recalibration processes, thus increasing their usability. Previous studies have suggested that ErrP signals are rather stable across recording sessions. Also, studies using protocols of serial stimuli presentation show that these potentials do not change significantly with the presentation rate. Here we complement these studies by analysing the decoding generalisation capabilities. Using data from monitoring experiments, we evaluate how much the performance degrades when tested in a condition different than the one the decoder was trained with. Moreover, we compare different spatial filtering techniques to see which preprocessing steps yield less-sensitive features for ErrP decoding.

show abstract

EEG-based decoding of error-related brain activity in a real-world driving task

Cited by 102 publications

References 30 publications

Plug-and-Play Supervisory Control Using Muscle and Brain Signals for Real-Time Gesture and Error Detection

Plug-and-Play Supervisory Control Using Muscle and Brain Signals for Real-Time Gesture and Error Detection

Real-time Detection of Driver’s Movement Intention in Response to Traffic Lights

Spatial filters yield stable features for error-related potentials across conditions

Contact Info

Product

Resources

About