Sleep disorder is a medical disease of the sleep patterns, which commonly suffered by the elderly. Sleep disorders diagnosis and treatment are considered to be challenging due to a time-consuming and inconvenient process for the patient. Moreover, the use of Polysomnography (PSG) in sleep disorder diagnosis is a high-cost process. Therefore, we propose an efficient classification method of sleep disorder by merely using electrocardiography (ECG) signals to simplify the sleep disorders diagnosis process. Different from many current related studies that applied a five-minute epoch to observe the main frequency band of the ECG signal, we perform a pre-processing technique that suitable for the 30-seconds epoch of the ECG signal. By this simplification, the proposed method has a low computational cost so that suitable to be implemented in an embedded hardware device. Structurally, the proposed method consists of five stages: (1) pre-processing, (2) spectral features extraction, (3) sleep stage detection using the Decision-Tree-Based Support Vector Machine (DTB-SVM), (4) assess the sleep quality features, and (5) sleep disorders classification using ensemble of bagged tree classifiers. We evaluate the effectiveness of the proposed method in the task of classifying the sleep disorders into four classes (insomnia, Sleep-Disordered Breathing (SDB), REM Behavior Disorder (RBD), and healthy subjects) from the 51 patients of the Cyclic Alternating Pattern (CAP) sleep data. Based on experimental results, the proposed method presents 84.01% of sensitivity, 94.17% of specificity, 86.27% of overall accuracy, and 0.70 of Cohen’s kappa. This result indicates that the proposed method able to reliably classify the sleep disorders merely using the 30-seconds epoch ECG in order to address the issue of a multichannel signal such as the PSG.

Emotions represent affective states that induce alterations in behavior and interactions within one's environment. An avenue for discerning human emotions lies in the realm of speech analysis. Empirical evidence indicates that 1.6 million Indonesian teenagers grapple with mental anxiety disorders, characterized by sensations of fear or ambiguous vigilance. This work endeavors to devise a tool for discerning an individual's emotional state through voice processing, focusing particularly on fear emotions stratified into three levels of intensity: low, medium, and high. The proposed system employs Gammatone-Frequency Cepstral Coefficients (GFCC) for feature extraction, leveraging the efficacy of its gamma filter in reducing noise. Furthermore, a Random Forest (RF) Classifier is integrated to facilitate the recognition of fear's emotional intensity in speech signals. The system is deployed on a Raspberry Pi 4B and establishes a Bluetooth connection using the RFCOMM communication protocol to an Android application, presenting the classification results. The outcomes reveal that the Signal-to-Noise Reduction achieved through GFCC extraction surpasses that of Mel-Frequency Cepstral Coefficients (MFCC). In terms of accuracy, the implemented recognition system for fear emotion levels, employing GFCC extraction and Random Forest Classifier, attains a commendable accuracy of 73.33 %.

The global prevalence of arm amputations, particularly transradial or below-elbow amputations, presents a pressing challenge due to the high costs of commercially available prosthetic hands. This issue is particularly acute in developing countries, where financial constraints hinder access to functional prosthetic solutions. Addressing this gap, this research introduces an innovative approach that merges electromyography (EMG) signal recognition and machine learning to develop an affordable and versatile bionic below-elbow prosthetic hand. The integration of EMG signals for control, coupled with machine learning algorithms, enables accurate classification of hand movements, resulting in enhanced movement precision and a broader range of gestures. This solution overcomes limitations in existing prosthetic hands, offering advanced EMG-based control and comprehensive movement capabilities, empowering users to perform intricate tasks and regain functional autonomy. The study undertakes a comprehensive analysis of the developed system's performance, encompassing hardware, software, and machine learning components. Results show promising accuracy levels for movement classification using the K-Nearest Neighbors (KNN) machine learning method. Execution time variations are observed among subjects and gestures, influenced by factors such as subject sensitivity and gesture complexity. Energy consumption analysis reveals disparities in power usage among different movements, emphasizing the impact of movement complexity on power requirements. A cost analysis demonstrates the affordability of the prototype, positioning it as a cost-effective alternative to existing prosthetic hand options. Comparing this solution to other research and commercial prosthetic hands highlights its uniqueness, particularly its emphasis on affordability.

COVID-19 (Coronavirus Disease 2019) is an infectious disease caused by the SARS-CoV-2 virus. This disease has spread worldwide since the beginning of 2020. Patients with this highly contagious disease generally experience only mild to moderate respiratory problems such as sore throat, cough, runny nose, fever, shortness of breath, and fatigue. However, some will become seriously ill and may cause severe respiratory distress or in severe cases multiple organ failure. Therefore, early identification of COVID-19 patients is very important. In this study, a disease detection system was created using an open dataset from COUGHVID which were contained the coughing sound of the Covid-19 disease. The implementation of the cough voice recognition system uses the K-Nearest Neighbor (K-NN) machine learning method and the Linear Predictive Coding (LPC) as method of extracting features from voice. The system was built using the Raspberry Pi 3 b+ microcontroller with microphone voice input and connected to a 3.5-inch LCD touchscreen display as the interface of the system device. The test uses a coughing sound as input through a microphone and processed by LPC feature extraction. At each running process, about 399 MB of memory is used from a total of 1 GB of memory. Meanwhile, the prediction of coughing sounds with the K-NN classification algorithm using 5 neighbors produces accuracy of 62% to predict disease.

Humanoid robots require advanced balance control systems to safely and efficiently operate in dynamic and unstructured environments. The research contribution is an IMU-only balance-detection framework introducing a quantitative Responsiveness Index (RI) to compare the responsiveness of Center of Mass (CoM) and model-based Zero Moment Point (ZMP) estimation. The proposed method integrates CoM and ZMP analyses derived exclusively from Inertial Measurement Unit (IMU) data and employs an Unscented Kalman Filter (UKF) with optimized parameters (?=0.10, ?=2, process noise covariance Q = 0.10, measurement noise covariance R = 0.10) obtained through sensitivity analysis across multiple motion states. This configuration effectively suppresses IMU noise and stabilizes attitude estimation on the DARwIn-OP3 humanoid platform. Experimental validation showed that CoM estimation achieved accuracies of 85.71 % for x, 97.85 % for y, and 99.96 % for z relative to CAD-based static references, while ZMP estimation reached 71.43 % for x and 94.89 % for y, and the Support Polygon (SP) area matched the model within ± 0.5 %. The proposed RI metric revealed that ZMP detected balance transitions approximately 20 % faster than CoM during dynamic walking and kicking, confirming its superior responsiveness. These findings demonstrate that IMU-derived, model-based ZMP estimation provides an effective, real-time indicator of humanoid balance changes, enabling future development of closed-loop control strategies to maintain ZMP within the SP for improved dynamic stability.

Stress is a state of emotional tension generated by a variety of factors such as work, study, family, and others. Stress can worsen and have an impact on health if it is not managed soon. Several studies have presented methods for detecting people's emotions through their voices. The goal of this study is to determine whether someone is stressed and how much stress he is under by listening to his voice. It is expected to be able to assess the amount of stress through sound utilizing MFCC feature extraction and artificial neural network machine learning. This system is powered by a Raspberry Pi 4 connected through Bluetooth to a microphone and an application on an Android phone. The smartphone application was designed to integrate with the embedded system and to display the prediction result. The dataset used in this research was SUSAS (Speech Under Simulated and Acute Stress) consisting of 1860 utterances. During the development of the artificial neural network model using 70% as training dataset, the accuracy only achieved 76%. However, the accuracy of the overall integrated system by utilizing 30 data taken from dataset reached 90%. Meanwhile, the test's average computing time is 2 seconds.