This study presents a deep learning model devoted to the analysis of swimming using a single Inertial Measurement Unit (IMU) attached to the sacrum. Gyroscope and accelerometer data were collected from 35 swimmers with various expertise levels during a protocol including the four swimming techniques. The proposed methodology took high inter- and intra-swimmer variability into account and was set up for the purpose of predicting eight swimming classes (the four swimming techniques, rest, wallpush, underwater, and turns) at four swimming velocities ranging from low to maximal. The overall F1-score of classification reached 0.96 with a temporal precision of 0.02 s. Lap times were directly computed from the classifier thanks to a high temporal precision and validated against a video gold standard. The mean absolute percentage error (MAPE) for this model against the video was 1.15%, 1%, and 4.07%, respectively, for starting lap times, middle lap times, and ending lap times. This model is a first step toward a powerful training assistant able to analyze swimmers with various levels of expertise in the context of in situ training monitoring.

This paper presents a novel sensor-to-segment calibration procedure for inertial sensor-based knee joint kinematics analysis during cycling. This procedure was designed to be feasible in-field, autonomously, and without any external operator or device. It combines a static standing up posture and a pedaling task. The main goal of this study was to assess the accuracy of the new sensor-to-segment calibration method (denoted as the ‘cycling’ method) by calculating errors in terms of body-segment orientations and 3D knee joint angles using inertial measurement unit (IMU)-based and optoelectronic-based motion capture. To do so, 14 participants were evaluated during pedaling motion at a workload of 100 W, which enabled comparisons of the cycling method with conventional calibration methods commonly employed in gait analysis. The accuracy of the cycling method was comparable to that of other methods concerning the knee flexion/extension angle, and did not exceed 3.8°. However, the cycling method presented the smallest errors for knee internal/external rotation (6.65 ± 1.94°) and abduction/adduction (5.92 ± 2.85°). This study demonstrated that a calibration method based on the completion of a pedaling task combined with a standing posture significantly improved the accuracy of 3D knee joint angle measurement when applied to cycling analysis.

Characterization of muscle mechanism through the torque-angle and torque–velocity relationships is critical for human movement evaluation and simulation. in vivo determination of these relationships through dynamometric measurements and modeling is based on physiological and mathematical aspects. However, no investigation regarding the effects of the mathematical model and the physiological parameters underneath these models was found. The purpose of the current study was to compare the capacity of various torque-angle and torque–velocity models to fit experimental dynamometric measurement of the elbow and provide meaningful mechanical and physiological information. Therefore, varying mathematical function and physiological muscle parameters from the literature were tested. While a quadratic torque-angle model seemed to increase predicted to measured elbow torque fitting, a new power-based torque–velocity parametric model gave meaningful physiological values to interpret with similar fitting results to a classical torque–velocity model. This model is of interest to extract modeling and clinical knowledge characterizing the mechanical behavior of such a joint.

This article focuses on the design of a sensor system for a non-planar surface, in particular a cylindrical shape, such as a kayak paddle. The main objective is to develop a piezoresistive sensor system to measure the pressure exerted by the hand on the shaft. The study begins with static characterization of the sensors, including dispersion analysis to assess their sensitivity, linearity and measurement range. A calibration process is carried out using a dedicated test bench, and an inverse viscoelastic model is used to establish an accurate relationship between the measured resistance and the corresponding pressure. The sensor system is connected to a data acquisition board equipped with an analog-to-digital converter (ADC) that enables the direct conversion of analog data into digital resistance values. Furthermore, Bluetooth Low Energy (BLE) wireless communication is employed to facilitate data transfer to a computer, enabling a detailed pressure mapping of the kayak paddle and real-time data collection. The calibrated sensors are then tested and validated on the kayak paddle, facilitating the mapping of pressure zones on the paddle surface. This mapping provides information for locating areas of high pressure exertion during kayaker movements.

The widespread adoption of instrumented textiles has made a significant impact in various domains, encompassing health monitoring, rehabilitation, biomechanics, and sports. This study specifically focuses on the development and evaluation of a smart garment that employs low-energy flexible sensors embedded within the fabric to effectively monitor upper body movements. These sensors utilize a piezoresistive polymer integrated into the garment and establish a connection with an electronic board for data acquisition. Wireless data transmission is achieved through the utilization of Bluetooth Low Energy (BLE) technology, with the garment showcasing an impressive average power consumption of approximately 10 μW. To ensure the sensor’s performance and reliability, a comprehensive characterization process is meticulously conducted utilizing a dedicated test bench. Furthermore, this study conducts a comparative analysis between two distinct estimators utilized for determining the flexion/extension angles of the upper body joint. The first estimator leverages a Nonlinear AutoRegressive exogenous (NARX) neural network model, while the second estimator employs a viscoelastic model. Through extensive evaluation, it becomes evident that the NARX neural network model outperforms the viscoelastic model, showcasing superior accuracy with a root-mean-square error of 4.85°. Consequently, the NARX neural network model emerges as the preferred option for accurately estimating the flexion/extension angles of the upper body joint.