Electroencephalogram-Based Brain-Computer Interface for Internet of Robotic Things

Motion intention detection is fundamental in the implementation of human-machine interfaces applied to assistive robots. In this paper, multiple machine learning techniques have been explored for creating upper limb motion prediction models, which generally depend on three factors: the signals collected from the user (such as kinematic or physiological), the extracted features and the selected algorithm. We explore the use of different features extracted from various signals when used to train multiple algorithms for the prediction of elbow flexion angle trajectories. The accuracy of the prediction was evaluated based on the mean velocity and peak amplitude of the trajectory, which are sufficient to fully define it. Results show that prediction accuracy when using solely physiological signals is low, however, when kinematic signals are included, it is largely improved. This suggests kinematic signals provide a reliable source of information for predicting elbow trajectories. Different models were trained using 10 algorithms. Regularization algorithms performed well in all conditions, whereas neural networks performed better when the most important features are selected. The extensive analysis provided in this study can be consulted to aid in the development of accurate upper limb motion intention detection models.

Section: Introductionmentioning

confidence: 99%

Elbow Motion Trajectory Prediction Using a Multi-Modal Wearable System: A Comparative Analysis of Machine Learning Techniques

Little

Pappachan

Yang

et al. 2021

“…Human intent has been decoded using technology such as functional magnetic resonance imaging, magnetoencephalography, functional near-infrared spectroscopy, and electroencephalography (EEG). EEG signals from the scalp using wet electrodes are widely used in communication [ 4 , 5 , 6 ], rehabilitation [ 7 , 8 ] due to these electrodes cost-effectiveness and high temporal resolution. However, conductive gels and glues are required to attach wet electrodes to the scalp [ 9 ], and the impedance of such gels and glues worsens over time [ 10 ], which makes it difficult to obtain stable measurements over a long period of time.…”

Section: Introductionmentioning

confidence: 99%

3D Printable Dry EEG Electrodes with Coiled-Spring Prongs

Kimura

Nakatani

Nishida

et al. 2020

Various dry electroencephalography (EEG) electrodes have been developed. Dry EEG electrodes need to be pressed onto the scalp; therefore, there is a tradeoff between keeping the contact impedance low and maintaining comfort. We propose an approach to solve this tradeoff through the printing of complex-shaped electrodes by using a stereolithography 3D printer. To show the feasibility of our approach, we fabricated electrodes that have flexible fingers (prongs) with springs. Although dry electrodes with flexible prongs have been proposed, a suitable spring constant has not been obtained. In this study, the spring constant of our electrodes was determined from a contact model between the electrodes and the scalp. The mechanical properties and reproductivity of the electrodes were found to be sufficient. Finally, we measured the alpha waves when a participant opened/closed his eyes by using our electrodes.

“…This becomes particularly problematic when only a few eras are available and artifacts such as blinking or movement are too frequent. Moreover, this approach is inappropriate for working with [ 5 ]: Continuous EEG and activity not blocked by events [ 6 ]; Long-range time correlations [ 7 ]; Real-time Brain to Computer (BCI) interface applications [ 8 , 9 ] for example speed control of Festo Robotino mobile robot using NeuroSky MindWave EEG headset based brain-computer interface [ 10 ] whether the BCI system designed for human-computer based control of IoT based robot (IoRT) [ 11 ]; Online mental health monitoring [ 12 ]. …”

Section: Introductionmentioning

confidence: 99%

“…Real-time Brain to Computer (BCI) interface applications [ 8 , 9 ] for example speed control of Festo Robotino mobile robot using NeuroSky MindWave EEG headset based brain-computer interface [ 10 ] whether the BCI system designed for human-computer based control of IoT based robot (IoRT) [ 11 ];…”

Section: Introductionmentioning

confidence: 99%

Determination of the Inaccuracies of Calculated EEG Indices

Borawski

Biercewicz

Duda

2020

The data obtained as a result of an EEG measurement are burdened with inaccuracies related to the measurement process itself and the need to remove recorded disturbances. The article presents an example of how to calculate the Approach-Withdraw Index (EEG-AW) and Memorization Index (MI) indices in such a way that their inaccuracy resulting from the removal of artifacts can be periodically calculated. This inaccuracy is expressed in terms of standard deviation. This allows you to determine the reliability of the obtained conclusions in the context of examining elements in a 2D computer game created in the Unity engine.