A Raspberry Pi-Based Traumatic Brain Injury Detection System for Single-Channel Electroencephalogram

Autonomous sleep tracking at home has become inevitable in today’s fast-paced world. A crucial aspect of addressing sleep-related issues involves accurately classifying sleep stages. This paper introduces a novel approach PSO–XGBoost, combining particle swarm optimisation (PSO) with extreme gradient boosting (XGBoost) to enhance the XGBoost model’s performance. Our model achieves improved overall accuracy and faster convergence by leveraging PSO to fine-tune hyperparameters. Our proposed model utilises features extracted from EEG signals, spanning time, frequency, and time–frequency domains. We employed the Pz-oz signal dataset from the sleep-EDF expanded repository for experimentation. Our model achieves impressive metrics through stratified-K-fold validation on ten selected subjects: 95.4% accuracy, 95.4% F1-score, 95.4% precision, and 94.3% recall. The experiment results demonstrate the effectiveness of our technique, showcasing an average accuracy of 95%, outperforming traditional machine learning classifications. The findings revealed that the feature-shifting approach supplements the classification outcome by 3 to 4 per cent. Moreover, our findings suggest that prefrontal EEG derivations are ideal options and could open up exciting possibilities for using wearable EEG devices in sleep monitoring. The ease of obtaining EEG signals with dry electrodes on the forehead enhances the feasibility of this application. Furthermore, the proposed method demonstrates computational efficiency and holds significant value for real-time sleep classification applications.

Section: Experiments 1: Classification Using Default Hyperparametersmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

An Autonomous Sleep-Stage Detection Technique in Disruptive Technology Environment

Radhakrishnan,

Ezra,

Jebadurai

et al. 2024

“…The system consists of a Digital-to-Analog Converter (DAC) in charge of converting the stored data to analog signals mimicking real EEG acquisition scenarios [ 20 ]. The DAC is the 12-bit MCP4725 chip with an Inter-Integrated Circuit (I2C) communication bus [ 21 ].…”

Section: Hardware Implementationmentioning

confidence: 99%

Investigation of EEG-Based Biometric Identification Using State-of-the-Art Neural Architectures on a Real-Time Raspberry Pi-Based System

Benomar

Cao

Vishwanath

et al. 2022

Self Cite

Despite the growing interest in the use of electroencephalogram (EEG) signals as a potential biometric for subject identification and the recent advances in the use of deep learning (DL) models to study neurological signals, such as electrocardiogram (ECG), electroencephalogram (EEG), electroretinogram (ERG), and electromyogram (EMG), there has been a lack of exploration in the use of state-of-the-art DL models for EEG-based subject identification tasks owing to the high variability in EEG features across sessions for an individual subject. In this paper, we explore the use of state-of-the-art DL models such as ResNet, Inception, and EEGNet to realize EEG-based biometrics on the BED dataset, which contains EEG recordings from 21 individuals. We obtain promising results with an accuracy of 63.21%, 70.18%, and 86.74% for Resnet, Inception, and EEGNet, respectively, while the previous best effort reported accuracy of 83.51%. We also demonstrate the capabilities of these models to perform EEG biometric tasks in real-time by developing a portable, low-cost, real-time Raspberry Pi-based system that integrates all the necessary steps of subject identification from the acquisition of the EEG signals to the prediction of identity while other existing systems incorporate only parts of the whole system.

“…The Raspberry Pi 4.0 is a high-performance 64-bit quad-core processor, up to 8 GB of RAM, dual-band 2.4/5.0 GHz wireless local area network (LAN), Bluetooth 5.0, Gigabit Ethernet, USB 3.0, and a sense HAT add-on which is attached on top of the Raspberry Pi via the 40 general-purpose input or output pins (which provide the data and power interface). It has several sensors and an 8 × 8 RGB (Red–Green–Blue) LED matrix display that can be used to visualize sensor states for multiple applications [ 47 , 48 , 49 , 50 ]. In the proposed design, the critical component, because of its value and relative complexity is the Raspberry Pi CPU.…”

Section: Case Study Quantum Jidokamentioning

confidence: 99%

Quantum JIDOKA. Integration of Quantum Simulation on a CNC Machine for In–Process Control Visualization

Villalba-Díez

Gutiérrez

Grijalvo

et al. 2021

With the advent of the Industry 4.0 paradigm, the possibilities of controlling manufacturing processes through the information provided by a network of sensors connected to work centers have expanded. Real-time monitoring of each parameter makes it possible to determine whether the values yielded by the corresponding sensor are in their normal operating range. In the interplay of the multitude of parameters, deterministic analysis quickly becomes intractable and one enters the realm of “uncertain knowledge”. Bayesian decision networks are a recognized tool to control the effects of conditional probabilities in such systems. However, determining whether a manufacturing process is out of range requires significant computation time for a decision network, thus delaying the triggering of a malfunction alarm. From its origins, JIDOKA was conceived as a means to provide mechanisms to facilitate real-time identification of malfunctions in any step of the process, so that the production line could be stopped, the cause of the disruption identified for resolution, and ultimately the number of defective parts minimized. Our hypothesis is that we can model the internal sensor network of a computer numerical control (CNC) machine with quantum simulations that show better performance than classical models based on decision networks. We show a successful test of our hypothesis by implementing a quantum digital twin that allows for the integration of quantum computing and Industry 4.0. This quantum digital twin simulates the intricate sensor network within a machine and permits, due to its high computational performance, to apply JIDOKA in real time within manufacturing processes.