1. Introduction
Because the human brain is a dynamic and complex structure, it has been studied for decades, and brain-computer interface systems have evolved to explore new ways of harnessing its power to improve human life.In particular, brain-computer interface systems aim to establish direct communication paths between the brain and external devices, bypassing the body’s more typical nerve and muscle pathways [1].Numerous studies have explored the potential of BCI systems in a variety of applications, including rehabilitation [2]navigation and robot control [3]environmental control [4]as well as games and entertainment [5].
Depending on the chosen experimental method and the expected neurophysiological activation patterns, multiple forms of task-relevant information can be retrieved from brain waves. Important examples include evoked potential (EP), steady-state evoked potential (SSEP) [6]event related potential (ERP) [7]and sensorimotor rhythms as motor imagery [8].
Brain-computer interfaces (BCIs) based on motor imagery, in which users imagine performing specific actions without actually performing them, have emerged as a promising approach to facilitate communication and control in people with movement disabilities [9] and universal applications.In stroke rehabilitation applications, MI-controlled robotic arms have been used to guide arm movements in patients undergoing stroke rehabilitation. [10]while virtual reality has been used in upper limb rehabilitation [11].Continuous game control via MI-BCI is now available [12]and specific feedback based on immersive virtual reality has been [13] Improve MI-BCI control.exist [14]designed an adaptive BCI speller based on motor imagery.
Additionally, as the number of smart devices in our homes continues to grow, so does the need for efficient and convenient control systems. In this direction, several methods have been proposed in the literature in recent years.A prototype of SSVEP-based BCI for home appliance control is presented in [15].Surface electromyography (sEMG) readings from the occipital region have been used to drive home automation systems [16]. “Neurophone” [17] Recognizing mental commands by gamma feature bands using trained hidden Markov models. The ‘BackHome’ system, powered by the P300 control interface, brings together a suite of services including smart home control, cognitive stimulation, online browsing, remote monitoring and home support tools to facilitate implementation in the home for users and caregivers without the need for specialized training. autonomy [18]. To the best of the authors’ knowledge, the potential of MI-BCI to provide an intuitive way to control home devices, especially for individuals with physical disabilities, has not been fully explored among the methods presented, and this paper aims to fill this gap.
MI-BCI based systems need to interpret the user’s intended actions. Such a task is accomplished by first classifying EEG signals and then converting them into commands. In recent years, several classifiers have emerged as popular techniques to address this challenge.These techniques include Bayesian methods [19]pattern matching [20]Neural Networks [21],Support Vector Machines [22]whitening technology based on Gram-Schmidt orthogonalization [23]linear discriminant analysis [24]. LDA is a supervised classification algorithm that has been widely and successfully applied to BCI problems due to its simplicity and high classification accuracy.
In addition, if the feature space dimension is large, spatial filters can be used to reduce the number of features to prevent the classifier from overfitting. The most straightforward approach is to manually select features from a priori data inspection. However, automation can be achieved through the use of statistical methods. A well-known method of extracting brain activity used in MI-BCIs is represented by Common Spatial Patterns (CSP) [25], a feature extraction technique that optimizes the separation of different signal classes.To improve the performance of the CSP algorithm in high-dimensional data settings, adding regularization terms to the CSP algorithm is an effective solution to improve robustness against noisy or incomplete data. [26]. The combination of these two well-known methods (i.e., LDA and regularized CSP) can make the classification of EEG signals more accurate and efficient.
That is why in this work, we propose a novel use of MI-based BCI system that utilizes LDA and RCSP to drive home automation systems. The entire software architecture is based on Konnex, a home automation standardized protocol that facilitates communication between hardware devices to control various home appliances in real time. Furthermore, the proposed framework enables simultaneous control of two different devices by providing real-time information about the device status.
Our results show that the proposed BCI system achieves good classification accuracy and good response time, indicating its application potential in home automation systems.
The paper is structured as follows: Section 2 introduces the participants, experiments, data acquisition, preprocessing and classification phases. Section 3 details the experimental part consisting of software and hardware architecture and the experimental results. Finally, section refsec:conc presents conclusions and future work.
4.in conclusion
In this work, we address the challenging problem of designing a MI-BCI based smart home interface to drive smart home appliances in real time. Our work shows that the problem can be solved using software and hardware architectures that are designed, implemented and thoroughly tested on different topics to show the effectiveness of the proposed solution. In particular, the integration of accessible programming tools (e.g. Node-RED) and reliable communication protocols (e.g. “KNX”) for hardware device setup and communication makes it possible to implement a control framework that drives two switching devices with sufficient accuracy. possible.
Further research should be conducted to account for daily fluctuations in EEG signals, and more work needs to be done to ensure more reliable performance. In particular, applying such systems to a wider range of users and focusing on long-term experiments may also be critical next steps. Furthermore, given the versatility of the proposed communication protocol, it can be tested on more devices, allowing the user to expand the set of possible devices that the user can effectively control.
However, despite the effectiveness of the proposed approach, and given the complexity of the field, additional challenges still need to be faced. Additional surveys should be conducted to reduce the time of the training phase. The current implementation also requires the efforts of prominent entities. In addition, from a control perspective, further improvements should be made in the direction of enhancing system output. In fact, users should be given more freedom of control to adjust more than two devices simultaneously. MI multi-class classification can provide significant help in this direction. From a technical perspective, further experiments are needed to understand the reliability of the method in research laboratories, under real-world conditions, and in a variety of disciplines. These experiments may introduce additional challenges that have not yet been considered, such as the impact that real-world environmental perturbations may have on such systems.
Source link