Objective: The diagnosis of Sleep Apnea-Hypopnea Syndrome (SAHS) holds significant importance for assessing sleep quality and treating sleep disorders. However, the detection of hypopnea events has not been given due emphasis, and the precise delineation of event boundaries is not straightforward. In this work, we introduce a novel deep learning model for the precise detection of obstructive sleep apnea and hypopnea events. Methods: Respiration-related signals, processed through a sliding window, serve as inputs to the model. Initially, multi-scale features are extracted using the Dilated Pyramid Convolution module, followed by an adaptive refinement of these features using the Frequency Enhanced Attention module. Finally, the Contextual Representation Learning module captures the temporal dependencies within the features. Results: The model was validated on two public datasets and one local dataset, achieving an accuracy of 84.4%, a precision of 66.3%, a recall of 84.5%, and an F1 score of 72.3% on the SHHS2 dataset. Conclusion and significance: We have achieved an automatic detection of both obstructive sleep apnea and hypopnea events with a granularity of one second. Our method offers certain advantages over other approaches, with the potential to assist in clinical diagnosis and to enable home-based respiratory monitoring.

This paper focuses on collecting a dataset that will be used for the development of biofeedback systems. The measurement setup included an IMU sensor and a motion capture system. We have collected recordings of hand movement in over 2000 dart throws. Our preliminary ML experiments employed a 1D convolutional neural network (1D-CNN) to classify dart throws based on throwing accuracy and throwing precision. Results indicate that the system can achieve up to 57

This paper contains an analysis of methods for person classification based on signals from wearable IMU sensors during sports. While this problem has been investigated in prior work, existing approaches have not addressed it within the context of few-shot or minimal-data scenarios. A few-shot scenario is especially useful as the main use case for person identification in sports systems is to be integrated into personalised biofeedback systems in sports. Such systems should provide personalised feedback that helps athletes learn faster. When introducing a new user, it is impractical to expect them to first collect many recordings. We demonstrate that the problem can be solved with over 90% accuracy in both open-set and closed-set scenarios using established methods. However, the challenge arises when applying few-shot methods, which do not require retraining the model to recognise new people. Most few-shot methods perform poorly due to feature extractors that learn dataset-specific representations, limiting their generalizability. To overcome this, we propose a combination of an unsupervised feature extractor and a prototypical network. This approach achieves 91.8% accuracy in the five-shot closed-set setting and 81.5% accuracy in the open-set setting, with a 99.6% rejection rate for unknown athletes.

Wireless communication in aquatic environments faces major challenges as the radio channel is temporarily unavailable due to submersion and the movement of swimmers. These interruptions can lead to sudden and prolonged packet loss, making real-time biofeedback difficult to achieve with standard communication protocols. This paper presents a real-time swimmer monitoring system built around a novel Block Selective Repeat (BSR) communication protocol specifically designed for environments with cyclic channel interruptions. BSR extends the traditional selective repeat model by introducing dynamic packet composition, cumulative acknowledgements with loss indicators and buffered retransmission. These features enable efficient and reliable transmission of sensor data, even when the connection is temporarily interrupted. The system was implemented using low-cost, commercially available Wi-Fi enabled microcontrollers and tested under real swimming conditions in four swimming styles. Experimental results show that BSR achieves 100% data recovery, with consistent low-latency performance across all swimming styles, including high-loss scenarios such as breaststroke and butterfly. The protocol also preserves the temporal structure of the movement data, enabling accurate recognition of movement patterns for real-time coaching and performance analysis. These findings demonstrate the suitability of BSR for biofeedback systems in swimming and other applications where intermittent wireless connectivity is a constraint.

This study explores the application of wireless wearable devices within the emerging domain of biomechanical feedback systems for sports and rehabilitation. A critical aspect of these systems is the need for real-time operation, where wearable devices must execute multiple processes concurrently while ensuring specific tasks are performed within precise time constraints. To address this challenge, we developed a specialized, lightweight periodic scheduler for microcontrollers. Extensive testing under various conditions demonstrated that sensor sampling frequencies of 650 Hz and 1750 Hz are achievable when utilizing one and 26 sensor samples per packet, respectively. Receiver delays were observed to be a few milliseconds or more, depending on the application scenario. These findings offer practical guidelines for developers and practitioners working with real-time biomechanical feedback systems. By optimizing sensor sampling frequencies and packet configurations, our approach enables more responsive and accurate feedback for athletes and patients, improving the reliability of motion analysis, rehabilitation monitoring, and training assessments. Additionally, we outline the limitations of such systems in terms of transmission delays and jitter, providing insights into their feasibility for different real-world applications.

Accurate, field-ready timing and motion capture are essential for assessing agility beyond the limits of manual stopwatches. We present a modular measurement system that fuses infrared (IR) optical gates for robust event detection with a trunk-worn inertial measurement unit (IMU) for kinematic profiling. Each sensing node is built on an Adafruit Feather M0 Wi-Fi microcontroller and communicates via UDP to a laptop server. Time alignment is accomplished without internet connectivity: the server establishes a relative epoch and executes a triple-handshake broadcast protocol, while timestamps are generated at the edge to avoid latency bias from transport or processing. Module- and device-level characterization shows that IR-receiver processing combined with interrupt service routine latency yields a per-event timestamp error of 0.54 ms +/- 0.14 ms (latency +/- uncertainty), and local clocks remain stable over the durations relevant to agility trials. In wireless operation, accepted synchronization attempts form tight response clusters in favorable RF conditions, whereas congested environments may require retries; for section times across different gates we therefore report a conservative inter-node uncertainty. End-to-end validation across laboratory, entry-hall, and gym venues using the Agility T-test confirms that total test time measured on the same start/finish gate remains below 1 ms error over 10-20 s trials. Synchronized IMU waveforms add explanatory value beyond total and split times by revealing braking, change-of-direction, and re-acceleration phases. The system provides a deployable workflow with substantially improved precision over manual timing. Future work will target more robust synchronization and expanded analytics, including automated phase detection, asymmetry indices, and optional integration with indoor positioning

The use of wearable devices in sport is steadily increasing. This is particularly true for the applications using the concept of the real-time biomechanical feedback. The paper investigates the problem of providing such feedback to the user via tactile actuators in aquatic environments, which poses an additional challenge. A waterproof wearable device with six tactile actuators, control circuit, microcontroller and wireless connectivity is developed. Two studies are carried out with it. The exploratory study provides a basic understanding of the human perception of tactile interfaces in an aquatic environment. The device is used by 34 participants in eight separate trials determined by three independent variables i. e. a setup, environment and activity. Participants wear the devices around their waist and on their head, in and out of the water, while moving and at rest. A usability study involving 51 participants tests the use of the device during an intense sport tests exercise in an aquatic environment to establish whether users can respond to commands from the device. The first study uses 20 different tactile symbols, each symbol using a unique combination of one or more of the device actuators. The second study uses six best performing symbols from the first study. The paper ends by presenting the results of the exploratory and usability study and discusses their findings.

The escalating importance of wearable devices in sports, particularly those incorporating real-time biomechanical feedback, necessitates innovative solutions for user engagement. This study addresses the challenge of delivering such feedback through haptic actuators in aquatic environments. A waterproof wearable device featuring six haptic actuators, a wirelessly connected control circuit, and a microcontroller was engineered for this purpose. An exploratory study, involving 34 participants, was conducted to comprehensively assess human perception of haptic interfaces. The study encompassed eight distinct tests, considering three independent variables: Placement, Environment, and Action. Tests were conducted with the device worn at the waist and head, both in and out of water, and during both movement and stationary states. Expert evaluations correspond to results and endorse both mounting positions, favoring head mounting. Each test session involved exposure to 20 unique symbols, with distinct actuators activated for each symbol. The outcomes of this exploratory study are presented and discussed, providing insights into the challenges and potentials of utilizing haptic actuators for real-time biomechanical feedback in aquatic environments.

AbstractThis review focuses on the usage of machine learning methods in sports. It closely follows the PRISMA framework for writing systematic reviews. We introduce the broader field of using sensor data for feedback in sport and cite similar reviews, that focus on other aspects of the field. With its focus on machine learning models that use signals from simple sensors, this review covers a very focused area that has not yet been covered by any other review. As described in problem definition, we use well-defined inclusion criteria, we have reviewed 72 papers. They present existing solutions, that use machine learning to extract useful information from data collected using various sensors in sports. To be included, papers had to use machine learning methods using data collected from sensors during sports, had to focus on sports-related applications and the result of machine learning had to be some information that can be used in real-time. We have found that the field is rapidly developing as 46 of the 72 included papers were from the last four years. Furthermore, we have found that the field is moving from using classical machine learning techniques to using deep learning. We analyze which data is used as input for machine learning, and we find that the most commonly used sensor is the accelerometer, closely followed by the gyroscope. The most common sensor platform is using a single wearable sensor, however, the studies that used deep learning, use multiple wearable sensors most often. Dataset sizes of sports papers are relatively small compared to other fields, but datasets are on average slightly larger in studies that use deep learning than in those that do not. We analyze the most common preprocessing methods and find that low-pass filtering and feature extraction are commonly used. We compare different machine learning models and the results of the studies that have tested multiple models on the same data, where we find that deep learning proved to be better than classical machine learning. Most studies show classification accuracy of over 90%, showing that machine learning is a useful tool for the researched problems. We end the review by researching how far the machine learning methods were implemented. Twenty of the included papers used their machine learning models in applications beyond a research paper and provided some sort of feedback back to athletes or coaches. After completing the review of the field, we propose a solution – a plan for future research. The proposed solution is to use a combination of best practices from the included paper and methods that we found are not yet implemented in the field of sports. We further elaborate, where we see the current state of the field. We conclude the article with short summary of the findings.

The integration of wearable devices into sports and rehabilitation is expanding, particularly in real-time biomechanical feedback (RTBF) applications. This study addresses the challenge of delivering effective RTBF in aquatic environments through haptic feedback. A waterproof wearable device with six vibrotactile actuators, a control circuit, and a microcontroller was developed and evaluated. An exploratory study involving 21 participants (12 males, 9 females) was conducted to assess human perception of haptic feedback. The study examined three independent variables: placement (waist or head), environment (in water or on land), and activity (stationary or in motion). Participants performed eight test scenarios involving 20 distinct haptic patterns. Results indicated that both mounting positions were effective, with differences in recognition rates across conditions and symbol complexities. The findings offer insights into the design and application of haptic RTBF systems in aquatic settings, providing a foundation for future developments in sports and rehabilitation technologies.

This paper presents the development of an intuitive user interface (UI) for a biomechanical feedback (BMF) system to improve the golf swing. The main idea behind the research is to develop an application that will allow golfers to learn correct techniques more efficiently. Our BMF system consists of a mountable device with an Arduino microcontroller and a BNO086 orientation sensor that transmits data at 200 Hz to provide real-time feedback on club orientation. The user interface was developed using the Unity engine to ensure compatibility with different platforms and incorporate semi-realistic graphics to enhance the user’s understanding. By allowing comparisons between excellent and sub-optimal swings of an individual player, it addresses the unique techniques of individual golfers. This focus on individual’s technique is a new approach, that we haven’t found to be used any other research project.

Obtaining an objective information about exercise technique, progress and fatigue management is very important in the sport of weightlifting. This information is usually acquired subjectively by an athlete or a coach. Our motivation is to obtain an objective feedback, in our case physical parameters like vertical acceleration, vertical velocity, height and power produced on the bar by the athlete. We used a wireless sensor device that consists of an orientation sensor, which wirelessly transmits the sensors signals to the user. We performed accuracy tests and validated our sensor device with the professional optical motion capture system from Qualisys. With the validated sensor device, we carried out several measurements of muscle snatch weightlifting exercise. By combining sensors signals and a video recording, we performed an in-depth analysis of the exercise. The results show that the system is functioning as expected and produces the desired results. Using such an assistance system in training helps the coach and the athlete to analyze and evaluate weightlifting technique objectively. The work presented here is used in the development of a standalone weightlifting training assistance system.

Wearable devices have become indispensable tools in everyday life and sports, providing users with information and feedback using various modalities, built-in sensors and algorithms. Biomechanical systems consist of four components: users, sensors, actuators, and processing devices. In this study, the focus is on the actuators. Biomechanical feedback systems and applications can provide information to the user through different modalities, such as visual, auditory, or haptic. We have developed a feedback device that uses haptic actuators and can be used in underwater environments. Our exploratory study has shown that athletes can perceive haptic actuators and understand simple commands even when they are underwater. In this continuation of the study, we focus on the swimmers' ability to not only sense the actuators, but also to understand the information and translate it into a change in movement. 51 young swimmers tested this device in our experiments. Results show that the information from the haptic actuators can be perceived during swimming and that users can follow the commands from the haptic feedback by changing their motion. In this study, this ability was demonstrated by changing the swimming technique.

This paper explores the use of real-time augmented haptic biofeedback to improve the swimming technique of young professional swimmers. The study employs a specially designed wearable device with haptic actuators and an integrated orientation sensor. The device measures and stores the maximum body rotation angles during each swim stroke and gives the swimmers real-time feedback on their performance so that they can react, if needed, by changing their swimming technique. Two young professional swimmers (male: 19years old, female: 18years old) took part in a series of tests, including one medium tempo baseline swimming, two medium tempo and one fast tempo swimming sessions with feedback, a symmetry test with feedback using a snorkel and several sprint runs with feedback in 25-m and 50-m pools. Haptic feedback triggered by predefined thresholds based on baseline measurements led to a measurable reduction in maximum body rotation angles. The results suggest that real-time haptic feedback could be an effective means of positively influencing swimming technique. This initial study should serve as a basis for further studies to refine the study design, device design and application of different feedback modalities, as well as to explore the applicability of the technology in different swimming scenarios.

This paper discusses the use of wireless wearable devices for use in an emerging field of biomechanical feedback systems in sports and rehabilitation. Timing is particularly critical for biomechanical feedback systems where all operations should be completed in real time. Wearable devices in such systems must perform multiple processes in parallel, ensuring that certain processes are performed at certain times and certain processes are performed within certain time limits. We solved this challenge by using a lightweight real-time operating system for the microcontroller. We have tested our device extensively under a variety of different conditions and scenarios. The results can be used as guidelines for the optimal setting of sensor sampling frequency and number of samples in a packet, to show the limitations of such systems in terms of transmission delays and jitter, or for any other use in real-time biomechanical feedback systems.

Sleep is regarded as essential for maintaining optimal health and well-being; however, the existing sleep studies are homogeneous and subjectively intrusive. And, there is no golden standard for clinical sleep quality evaluation. Some researchers have explored sleep structure (sleep staging) as a crucial basis for sleep quality evaluation. Therefore, the paper proposes an automatic sleep structure recognition and analysis system to provide a novel approach to sleep quality evaluation. A deep convolutional neural network is applied to sleep structure recognition. It mainly includes data preprocessing, feature extraction, feature fusion and prediction. A sleep structure visualization platform with user-friendly interfaces and images is implemented to visualize and analyze the sleep structure. The method is evaluated on a both public and private dataset. The results show its good performance and a potential lore clinical application.

Sensors and smart equipment are frequently used in biomechanical systems and applications in sports and rehabilitation to measure various physical quantities. Various sensors, measuring different parameters, can produce a large amount of data at high speeds and volumes that must be stored for real-time or post-processing and analysis. In addition to sensor data, metadata is an important component and can vary between biomechanical applications. Currently however, each application typically has its own unique data flow and storage solution. In this research, we present a universal data model solution that can be applied to any sensor-based biomechanical application in sport and physical rehabilitation. Our proposed cloud platform architecture allows for the manipulation of sensor data and metadata using a combination of Big Data and conventional techniques. The main idea of this research is to develop a platform that allows a universal way for any biomechanical application to handle its data regardless of the type of data and metadata. This is achieved by creating a universal data model, and implementing this data model in a generalized architecture using a graph database. We demonstrate the benefits of this approach using examples from existing biomechanical systems and describe the development of the cloud platform architecture and the underlying data model. We also provide an example of the use of this platform in a sport shooting application. This approach is unique in that it allows data from different sources and applications to be stored and processed using the same procedures and techniques, facilitating data analysis and application development. We envision this system will expand to multiple different biomechanical applications in the future. We expect that in time, the ability to compare various data and store different biomechanical datasets will become necessity. With the advantages of modern recommender systems and utilization of artificial intelligence, huge amounts of relevant and well-prepared data with useful metadata are required thus having such system is an important advantage for future biomechanical systems development. With the increase of people's awareness and usage of devices that increase well-being and quality of life, presented platform and similar systems will play a pivotal role in shaping the future lifestyle.

Convolutional neural network (CNN)-based methods facilitate data classification but sacrifice physical interpretability due to the complex model architecture and tight inferring integration. The interpretability requirement of our prior CNN-based golf classifier motivates us to explain the performance of the predictions and to discover the class-discriminative, significant regions of interest within the golf swings as well. This can be done by casting the 2D Guided Grad-CAMs to a 1D generalization, which is presented in our current research. We then perform the visualization by inspecting the golf predictions and the involved golf dataset using such a custom 1D Guided Grad-CAM, highlight class-discriminative, significant regions of interest, and finally attempt to present potential interpretations. Specifically, we investigate the attention performance and the corresponding potential attributions by visualizing and by evaluating the predictions given by the classifier and the golf swings from five perspectives, including attention consistency within particular classes, the inspections of misclassified swings, Guided Grad-CAM visualizations at different layers, and the attention shift with respect to temporal resolutions and with respect to sensor usages. We conclude that our visual inspections explain our previous classification performance, that the class-discriminative, significant features can be captured, and that every single prediction has its reasonable interpretation, in terms of the comprehensive experiments. Such exploration can provide a potential possibility of associating the critical regions and features with the physical movements of golf players, which can possibly contribute to golf training. Relevant code files are available at https://github.com/92xianshen/golf-guided-gradcam .

The influence of joint motion on punch efficiency before impact is still understudied. The same applies to the relationship between the kinematic and temporal parameters of a reverse punch (RP) that determines a score. Therefore, the aim of this study was to investigate if the exclusion or inclusion of body segments affects the acceleration, velocity, rotation angle, and timeline of execution, and to examine the correlation between these quantities. Seven elite male competitors—senior European and World Championship medalists—participated in the in-field testing. Quantities were acquired in the developmental phase of RP through three modalities of execution. Synchronized real-time data were obtained using combined multimodal sensors and camera fusion. The main findings of the study have highlighted the significant differences in the temporal and kinematic variables of RP that arise from the modality of execution. Large and medium correlation coefficients were obtained between the examined variables of body and hand. In conclusion, the results show that measured parameters are affected by segmental body activation. Moreover, their interdependence influences punch execution. The presented interdisciplinary approach provides insightful feedback for: (i) development of reliable and easy-to-use technical solutions in combat sports monitoring; and (ii) improvements in karate training.

The use of high-tech wearable devices and real-time biomechanical feedback (RTBF) is widespread in modern sports and physical rehabilitation. While the use of kinematic sensors for RTBF is well established, the question of the most efficient and appropriate way to present feedback information to the user remains largely unanswered. The haptic modality has only been used in a limited number of relatively simple studies and applications, and it has never been studied in an aquatic environment. In this study, the main motivation was to design, develop, implement, and test an RTBF system for sports and physical rehabilitation with haptic actuators suitable for use in aquatic environments. The developed miniature remote-controlled device with vibration motors as actuators was tested to determine how people perceive the haptic modality in and out of water. The results of the exploratory study with 34 participants suggest that the feedback device can be further developed for future studies. The results show that both simple and complex haptic stimuli can be understood by athletes both outdoors and in water, as well as while standing and performing simple physical activities. This study tested the use of both head and waist mounted haptic actuators, with both showing similar and promising results. The results of this study provide evidence that haptic feedback can be perceived in water, highlighting the potential for real-time haptic interventions in aquatic environments. In summary, this study represents a significant contribution to the evolving field of RTBF in modern sport training and physical rehabilitation. The development of a haptic feedback device that can be worn during aquatic and other activities is a significant advance in the field of feedback actuators. This study provides a foundation for future usability studies and the development of haptic RTBF applications in both aquatic and outdoor environments.