Motion Data Research Articles

An enhanced hybrid scheme for ship roll prediction using support vector regression and TVF-EMD

In order to represent the complex characteristics of ship roll motion such as nonlinearity, uncertainty, and time-varying dynamics thus improving the ship roll prediction accuracy, an enhanced hybrid prediction scheme is proposed by using improved swarm intelligent optimization, time series decomposition, and machine learning. Firstly, the novel time-varying filtering-based empirical mode decomposition (TVF-EMD) is employed to decompose the ship roll motion data into multiple mode components with dynamic time-varying characteristics, which immunes of the mode mixing and intermittency problems of traditional empirical mode decomposition (EMD). Then, the support vector regression is utilized to train and predict mode components. Finally, the predicted results of individual components are reconstructed to achieve the final ship roll prediction angles. To avoid the unfavorable influence of manual parameter selection for TVF-EMD, an improved black widow optimization algorithm is employed to optimize the parameter configuration. The feasibility and effectiveness of the enhanced hybrid model are validated by ship roll prediction simulation based on the measured data of M.V. YuKun at sea. The experimental results show that the enhanced hybrid scheme can improve the prediction accuracy of ship roll motion and outrank those by using methods of EMD, ensemble empirical mode decomposition, and variational mode decomposition.

Aerodynamic interactions of blunt bodies free-flying in hypersonic flow

This paper takes a new look at how the aerodynamic interactions of multiple bodies in high-speed flow affect their motion behaviors. The influence of the body shape and orientation on aerodynamic and stability behavior in the case of shock–shock and wake–shock interactions is the focus of this publication. Experiments were performed in the hypersonic wind tunnel H2K at the German Aerospace Center (DLR) in Cologne. Free-flight tests with tandem arrangements of spheres and cubes were performed with a synchronized dropping of both objects at various initial conditions of relative streamwise and vertical distance as well as pitch angle. A high-speed stereo-tracking captured the model motions during free-flight, and high-speed schlieren videography provided documentation of the flow topology. Based on the measured 6-degrees-of-freedom (6DoF) motion data, aerodynamic coefficients were determined. As a result, the final lateral velocity of trailing cubes is found to be many times greater than that of spheres regarding shock-wave surfing. For rotating cubes, the results showed that stable shock-wave surfing can become possible over an increasingly wide range of initial positions. This study has identified that the trailing drag coefficient of two axially aligned objects varies strongly with their relative streamwise distance. Furthermore, it was shown that the wake is a region of stability for downstream objects.Graphical abstract

BACK-to-MOVE: Machine learning and computer vision model automating clinical classification of non-specific low back pain for personalised management.

Low back pain (LBP) is a major global disability contributor with profound health and socio-economic implications. The predominant form is non-specific LBP (NSLBP), lacking treatable pathology. Active physical interventions tailored to individual needs and capabilities are crucial for its management. However, the intricate nature of NSLBP and complexity of clinical classification systems necessitating extensive clinical training, hinder customised treatment access. Recent advancements in machine learning and computer vision demonstrate promise in characterising NSLBP altered movement patters through wearable sensors and optical motion capture. This study aimed to develop and evaluate a machine learning model (i.e., 'BACK-to-MOVE') for NSLBP classification trained with expert clinical classification, spinal motion data from a standard video alongside patient-reported outcome measures (PROMs). Synchronised video and three-dimensional (3D) motion data was collected during forward spinal flexion from 83 NSLBP patients. Two physiotherapists independently classified them as motor control impairment (MCI) or movement impairment (MI), with conflicts resolved by a third expert. The Convolutional Neural Networks (CNNs) architecture, HigherHRNet, was chosen for effective pose estimation from video data. The model was validated against 3D motion data (subset of 62) and trained on the freely available MS-COCO dataset for feature extraction. The Back-to-Move classifier underwent fine-tuning through feed-forward neural networks using labelled examples from the training dataset. Evaluation utilised 5-fold cross-validation to assess accuracy, specificity, sensitivity, and F1 measure. Pose estimation's Mean Square Error of 0.35 degrees against 3D motion data demonstrated strong criterion validity. Back-to-Move proficiently differentiated MI and MCI classes, yielding 93.98% accuracy, 96.49% sensitivity (MI detection), 88.46% specificity (MCI detection), and an F1 measure of .957. Incorporating PROMs curtailed classifier performance (accuracy: 68.67%, sensitivity: 91.23%, specificity: 18.52%, F1: .800). This study is the first to demonstrate automated clinical classification of NSLBP using computer vision and machine learning with standard video data, achieving accuracy comparable to expert consensus. Automated classification of NSLBP based on altered movement patters video-recorded during routine clinical examination could expedite personalised NSLBP rehabilitation management, circumventing existing healthcare constraints. This advancement holds significant promise for patients and healthcare services alike.

Tibial strains are sensitive to speed, but not grade, perturbations during running.

A fatigue-failure process is hypothesized to govern the development of tibial stress fractures, where bone damage is highly dependent on the peak strain magnitude. To date, much of the work examining tibial strains during running has ignored uphill and downhill running despite the prevalence of this terrain. This study examined the sensitivity of tibial strains to changes in running grade and speed using a combined musculoskeletal-finite element modeling routine. Seventeen participants ran on a treadmill at ±10°, ±5°, and 0°; at each grade, participants ran at 3.33 ms-1 and at a grade-adjusted speed -2.50 and 4.17 ms-1 for uphill and downhill grades, respectively. Force and motion data were recorded in each grade and speed combination. Muscle and joint contact forces were estimated using inverse-dynamics-based static optimization. These forces were applied to a participant-adjusted finite element model of the tibia. None of the strain variables (50th and 95th percentile strain and strained volume ≥4000 μɛ) differed as a function of running grade; however, all strain variables were sensitive to running speed (F ≥ 9.59, p ≤ 0.03). In particular, a 1 ms-1 increase in speed resulted in a 9% (≈260 μɛ) and 155% (≈ 600 mm3) increase in peak strain and strained volume, respectively. Overall, these findings suggest that faster running speeds, but not changes in running grade, may be more deleterious to the tibia. • High-magnitude tibial strains that are implicated in stress-fracture development were sensitive to changes in running speed but not grade. • The majority of the variance in measures of tibial strain was accounted for by individual subject variation, reinforcing the importance of inherent musculoskeletal properties in determining the bone strain environment.

HME-KG: A method of constructing the human motion encoding knowledge graph based on a hierarchical motion model

The diversity, infinity, and nonuniform description of human motion make it challenging for computers to understand human activities. To explore and reuse captured human motion data, this work defines a more comprehensive hierarchical theoretical model of human motion and proposes a standard human posture encoding scheme. We construct a domain knowledge graph (DKG) named the human motion encoding knowledge graph (HME-KG) based on posture codes and action labels. Community detection, similarity analysis, and centrality analysis are used to explore the potential value of motion data. This paper conducts an evaluation and visualization of HME-KG.

Immature characteristics of the East Anatolian Fault Zone from SAR, GNSS and strong motion data of the 2023 Türkiye–Syria earthquake doublet

On February 6, 2023, an Mw 7.9 earthquake occurred in the western section of the East Anatolia Fault Zone (EAFZ). It was subsequently followed by an Mw 7.7 earthquake on the northern branch of the EAFZ, known as the Sürgü Fault Zone. Coseismic deformation fields were derived for these earthquakes using joint evaluation of near-field strong motion data, Global Navigation Satellite System data, and Synthetic Aperture Radar datasets. The coseismic slip distribution model was determined through the joint kinematic finite fault inversion. The Mw 7.9 earthquake was a left-lateral strike-slip event, predominantly occurring at depths up to 20 km. The earthquake displayed three distinct asperities that correlate well with bends and stepovers along the EAFZ. The Mw 7.7 earthquake also exhibited left-lateral strike-slip characteristics, with a major asperity along the Çardak Fault featuring a maximum slip of approximately 9.5 m at depths between 0 and 24 km. The occurrence of this unanticipated large Mw 7.9 catastrophic seismic event on a fault with low-intermediate structural maturity is noteworthy. In the vicinity of immature faults with multiple jogs, stress tends to accumulate at barrier locations. When the accumulated stress near several adjacent barriers reaches a certain threshold, it may result in the transformation of multiple barriers into asperities, triggering cascading ruptures.

Reconstruction of a Long-Term, Reach-Scale Sediment Budget Using Lateral Channel Movement Data as a Proxy: A Case Study on the Lowland Section of the Tisza River, Hungary

Humans have influenced the sediment transport of rivers on a centurial scale. Our goal was to use the rate of lateral channel processes as a proxy to reconstruct sediment budget (SB) changes of a lowland river (Middle Tisza, Hungary) on a historical scale (1838–2017). The gross sediment budget (GSB) refers to the total area of eroded and accumulated surfaces, and the net sediment budget (NSB) indicates the sediment sink or source characteristics. At the beginning (1838–1890), the artificial cut-offs increased the slope and channel erosion, but the eroded sediment deposited in the oxbows, so the reach acted as a sediment sink (NSB: +0.1–0.8 m2/m/y). Then (1890–1929), a quasi-equilibrium state developed (NSB: −0.2 m2/m/y to +0.4 m2/m/y). Later (1929–1976), the bank protections impeded lateral erosion, so the system became a sediment sink again (NSB: +0.1–0.7 m2/m/y). Finally (1976–2017), the erosional processes accelerated due to dam construction and revetment collapses, and now the river is a sediment source (NSB: −0.03 to −0.08 m2/m/y). This study proved that (1) the actual SB could not be projected in the long term, as it was heavily modified, and (2) lateral channel changes could be used as a proxy to estimate long-term SB.

Markerless 3D kinematics and force estimation in cheetahs

The complex dynamics of animal manoeuvrability in the wild is extremely challenging to study. The cheetah (Acinonyx jubatus) is a perfect example: despite great interest in its unmatched speed and manoeuvrability, obtaining complete whole-body motion data from these animals remains an unsolved problem. This is especially difficult in wild cheetahs, where it is essential that the methods used are remote and do not constrain the animal’s motion. In this work, we use data obtained from cheetahs in the wild to present a trajectory optimisation approach for estimating the 3D kinematics and joint torques of subjects remotely. We call this approach kinetic full trajectory estimation (K-FTE). We validate the method on a dataset comprising synchronised video and force plate data. We are able to reconstruct the 3D kinematics with an average reprojection error of 17.69 pixels (62.94% PCK using the nose-to-eye(s) length segment as a threshold), while the estimates produce an average root-mean-square error of 171.3N (≈\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\approx$$\\end{document} 17.16% of peak force during stride) for the estimated ground reaction force when compared against the force plate data. While the joint torques cannot be directly validated against ground truth data, as no such data is available for cheetahs, the estimated torques agree with previous studies of quadrupeds in controlled settings. These results will enable deeper insight into the study of animal locomotion in a more natural environment for both biologists and roboticists.

Fall-risk assessment of aged workers using wearable inertial measurement units based on machine learning

Falls are the major cause of fatal and nonfatal incidents among aged workers. Estimating individual fall risk is crucial for occupational protection and public health. This study aimed to develop a fall-risk assessment method based on Machine Learning (ML) and Inertial Measurement Units (IMUs). A total of 28 aged workers (60 to 80 years old) participated in this study, recruited from community sanitation workers and janitors. They were categorized into high and low fall-risk groups based on functional gait assessment. Each participant performed 5 sets of motion experiments in the lab, covering 6 daily motions. We gathered their kinematic data with IMUs for training ML-based fall-risk assessment models. Five classic ML classifiers were optimized using grid search, and a Convolutional Neural Network and Long Short-Term Memory (CNN-LSTM) hybrid deep neural network was built to recognize walking motion from kinematic data for input into the fall-risk assessment model. The results demonstrated that ML classifiers trained with walking motion data exhibited superior performance. Cross-validation revealed that the optimized Random Forest classifier, with IMU on the right upper leg, reached 91.3 % accuracy. Additionally, the CNN-LSTM models were evaluated using F1-score, which can balance accuracy and coverage of model performance evaluation. The CNN-LSTM achieved a maximum F1-score of 95.18 %, proving their effectiveness in extracting walking motion data. These findings indicate that the developed IMU-based ML method exhibits excellent performance in assessing fall risk, potentially allowing aged workers to assess their fall risk before clinical diagnosis by embedding it in wearable IMUs.

Satellite group infall into the Milky Way: Exploring the Crater-Leo case with new HST proper motions

Within Lambda cold dark matter (Lambda CDM) simulations, Milky Way-like galaxies accrete some of their satellite galaxies in groups of 3-5 members rather than individually, and this has been suggested as a possible mechanism driving the formation of satellite planes. Objects accreted in groups are expected to share similar specific total energy and angular momentum, and to also have identical orbital planes and directions. Looking at observations of Milky Way satellites, the dwarf galaxies Leo II, IV, V, and Crater II, and the star cluster Crater 1 were proposed to be a vestige of group infall. The suggested `Crater-Leo group' shows a coherent distance gradient and all these objects align along a great circle on the sky. We used proper motion data to investigate whether the phase-space distribution of the members of the proposed group are indeed consistent with group infall. To further investigate this possibility, we used Gaia Data Release 3 (DR3) and new Hubble Space Telescope (HST) proper motions -- namely, $( alpha* 0.0523)$ mas $ for Leo IV and $( alpha* 0.1704)$ mas $ for Leo V -- to derive accurate orbital properties for the proposed group objects. In addition, we explored other possible members of this putative association. par Leo II, Leo IV, and Crater 1 show orbital properties consistent with those we predict from assuming group infall. However, our results suggest that Crater II was not accreted with the rest of the objects. If confirmed with increasingly accurate proper motions in the future, the Crater-Leo objects would appear to constitute the first identified case of a cosmologically expected, typical group infall event, as opposed to the highly hierarchical Magellanic Cloud system.

LSTM-based graph attention network for vehicle trajectory prediction

Vehicle Trajectory Prediction (VTP) is one of the key technologies for autonomous driving, which can improve the safety and collaboration of the autonomous driving system. The interaction behavior among vehicles in reality has an impact on VTP. However, many methods ignore the interaction among vehicles, which results in limited accuracy of prediction results. Therefore, we propose a Long Short-Term Memory (LSTM)-based Graph Attention Network (GAT) method for VTP, which encodes vehicle trajectory information with LSTM networks and represents vehicle interactions with GAT. Firstly, in order to capture the temporal relationship between positions and consider their influence, we use LSTM model to encode the position data. Meanwhile, to comprehensively model vehicle motion and use multidimensional feature representation, we employ another LSTM model to encode the motion data, including position, velocity and acceleration. Secondly, to learn distinct feature representation, we use one GAT module to process the LSTM position encoding features for capturing spatial relationships of position information. Another GAT module is employed to process the LSTM motion encoding features for fully considering multidimensional motion dynamics and spatial–temporal dependencies. Finally, the LSTM decoder receives all features and predicts the vehicle trajectory. The experimental results show that the proposed method demonstrates superior predictive performance by using the Next Generation Simulation (NGSIM) dataset.

Characterization of an Algorithm for Autonomous, Closed-Loop Neuromodulation During Motor Rehabilitation.

Recent evidence demonstrates that manually triggered vagus nerve stimulation (VNS) combined with rehabilitation leads to increased recovery of upper limb motor function after stroke. This approach is premised on studies demonstrating that the timing of stimulation relative to movements is a key determinant in the effectiveness of this approach. The overall goal of the study was to identify an algorithm that could be used to automatically trigger VNS on the best movements during rehabilitative exercises while maintaining a desired interval between stimulations to reduce the burden of manual stimulation triggering. To develop the algorithm, we analyzed movement data collected from patients with a history of neurological injury. We applied 3 different algorithms to the signal, analyzed their triggering choices, and then validated the best algorithm by comparing triggering choices to those selected by a therapist delivering VNS therapy. The dynamic algorithm triggered above the 95th percentile of maximum movement at a rate of 5.09 (interquartile range [IQR] = 0.74) triggers per minute. The periodic algorithm produces stimulation at set intervals but low movement selectivity (34.05%, IQR = 7.47), while the static threshold algorithm produces long interstimulus intervals (27.16 ± 2.01 seconds) with selectivity of 64.49% (IQR = 25.38). On average, the dynamic algorithm selects movements that are 54 ± 3% larger than therapist-selected movements. This study shows that a dynamic algorithm is an effective strategy to trigger VNS during the best movements at a reliable triggering rate.

Integrating human-centric simulations in educational production lines: advancing ergonomics for industry 5.0 applications

This research in the Industry 5.0 field focuses on a human-centered simulation of the FAS200 SMC educational production line, utilizing Tecnomatix Process Simulate Human software for developing a virtual human environment. A key aspect of this study is the integration of inertial sensors, enhancing the accuracy and depth of ergonomic analysis. These sensors play a pivotal role in capturing precise human movement data, crucial for ergonomic assessments. Adopting a defined working methodology, the study extensively employs the RULA method to evaluate operator postures in the production line. This approach has led to significant ergonomic improvements, evidenced by a 40 percent reduction in the RULA index at each workstation. The integration of inertial sensors has been instrumental in achieving these results, providing detailed insights into human movements and interactions with the production environment. The research transcends traditional ergonomic assessments by incorporating a new human-centered approach, emphasizing the well-being of individuals working alongside machines. This approach, bolstered by the use of inertial sensors, marks a significant advancement in ergonomic studies, aligning with the principles of Industry 5.0. The findings hold substantial potential for application in industrial settings, signaling a shift towards more human-friendly and efficient industrial practices.

Motion Capture Technology in Sports Scenarios: A Survey.

Motion capture technology plays a crucial role in optimizing athletes' skills, techniques, and strategies by providing detailed feedback on motion data. This article presents a comprehensive survey aimed at guiding researchers in selecting the most suitable motion capture technology for sports science investigations. By comparing and analyzing the characters and applications of different motion capture technologies in sports scenarios, it is observed that cinematography motion capture technology remains the gold standard in biomechanical analysis and continues to dominate sports research applications. Wearable sensor-based motion capture technology has gained significant traction in specialized areas such as winter sports, owing to its reliable system performance. Computer vision-based motion capture technology has made significant advancements in recognition accuracy and system reliability, enabling its application in various sports scenarios, from single-person technique analysis to multi-person tactical analysis. Moreover, the emerging field of multimodal motion capture technology, which harmonizes data from various sources with the integration of artificial intelligence, has proven to be a robust research method for complex scenarios. A comprehensive review of the literature from the past 10 years underscores the increasing significance of motion capture technology in sports, with a notable shift from laboratory research to practical training applications on sports fields. Future developments in this field should prioritize research and technological advancements that cater to practical sports scenarios, addressing challenges such as occlusion, outdoor capture, and real-time feedback.

Urban perception by using eye movement data on street view images

Urban perception by using eye movement data on street view images

Empirical models for Fourier amplitude spectrum of ground-motion calibrated on data from the Iranian plateau

AbstractGround-motion models (GMMs) are frequently used in engineering seismology to estimate ground motion intensities. The majority of GMMs predict the response spectral ordinates (such as spectral acceleration) of a single-degree-of-freedom oscillator because of their common application in engineering design practices. Response spectra show how an idealized structure reacts to applied ground motion; however, they do not necessarily represent the physics of ground motion. The functional forms of the response spectra GMMs are built around ideas taken from the Fourier spectral concept. Assuming the validity of Fourier spectral concepts in the response spectral domain could cause physically inexplainable effects. In this study, using a mixed-effects regression technique, we introduce four models capable of predicting the Fourier amplitude spectrum that investigates the impact of incorporating random-effect event and station terms and variations in using a mixed-effects regression technique in one or two steps using truncated dataset or all data (nontruncated dataset). All data consists of 2581 three-component strong ground motion data resulting from 424 events with magnitude ranging from 4.0 up to 7.4, from 1976 to 2020, and 706 stations. The truncated dataset’s records, events, and stations are reduced to 2071, 408, and 636, respectively. As part of this study, we develop GMMs to predict the Fourier amplitude spectrum for the Iranian plateau within the frequency range of 0.3–30 Hz. We adopted simple, functional forms for four models, and we included a limited number of predictors, namely Mw (moment magnitude), Rjb (Joyner–Boore distance), and VS30 (time-averaged shear-wave velocity in the top 30 m). Due to statistical analyses, the style-of-faulting term was excluded from the final functional forms. The robustness of the derived models is indicated by unbiased residual variation with predictor variables.

Salmon shark seasonal site fidelity demonstrates the influence of scale on identifying potential high-use areas and vulnerabilities

Considering habitat use throughout the whole range of a highly mobile marine species is necessary to understand life history, identify vulnerabilities, and inform effective management. We used satellite tagging data from 128 adult female salmon sharks Lamna ditropis to identify seasonal hotspots of activity in an extended California Current region (ECCR; encompassing the California Current Large Marine Ecosystem), an area far away from their well-described primary habitat in the Alaska Downwelling Region where they have been documented, but whose utility has been poorly understood. Tag track durations had a mean of 447.7 ± 381 d, and 88 sharks (68.8%) visited the ECCR, comprising 33.6% of 28019 total daily Argos detections. Tracking data revealed that the timing and duration of migrations to the ECCR varied, but salmon shark distribution within the ECCR displayed consistent latitudinal shifts in accordance with regional oceanographic seasons. High site fidelity across multi-year tracks to high-productivity features, such as sea banks, and previously published knowledge of salmon shark life history suggest that the ECCR provides important foraging habitat which may be linked to reproductive success. The data reveal high overlap of salmon shark distribution with cumulative fishing effort collected by Global Fishing Watch for 2012-2019, particularly around seasonal hotspots, suggesting that female salmon sharks might be at risk of fisheries encounters. Collectively, our findings emphasize the importance of the ECCR in salmon shark life history and demonstrate the influence of spatial and temporal scale on interpretation of large movement data sets and identification of critical habitat outside of well-studied regions.

Mobility census for monitoring rapid urban development.

Monitoring urban structure and development requires high-quality data at high spatio-temporal resolution. While traditional censuses have provided foundational insights into demographic and socio-economic aspects of urban life, their pace may not always align with the pace of urban development. To complement these traditional methods, we explore the potential of analysing alternative big-data sources, such as human mobility data. However, these often noisy and unstructured big data pose new challenges. Here, we propose a method to extract meaningful explanatory variables and classifications from such data. Using movement data from Beijing, which are produced as a by-product of mobile communication, we show that meaningful features can be extracted, revealing, for example, the emergence and absorption of subcentres. This method allows the analysis of urban dynamics at a high-spatial resolution (here 500 m) and near real-time frequency, and high computational efficiency, which is especially suitable for tracing event-driven mobility changes and their impact on urban structures.

Application of VR Motion Intelligent Capture Based on DLPMA Algorithm in Sports Training

With the rapid development of Virtual Reality (VR) technology, its application in the field of sports training is also receiving increasing attention. This study applies the Distance Likelihood Based Probabilistic Model Averaging (DLPMA) algorithm to the VR motion intelligent capture system, aiming to provide an efficient and accurate motion data collection method to improve existing sports training methods. Introduced the design and implementation of a VR motion intelligent capture system based on DLPMA algorithm, and applied it to sports training. By conducting comparative experiments with traditional training methods, the advantages of the system in motion capture accuracy, real-time performance, and user experience are verified. The research results indicate that the system can accurately capture the movements of athletes and provide timely feedback to users, providing an effective auxiliary means for sports training. Although the system has shown good performance in sports training, there are still some limitations. Future research can further optimize algorithms, enhance system stability and flexibility, to meet a wider range of sports training needs.

The Response of Taupō Volcano to the M7.8 Kaikōura Earthquake

AbstractSeveral studies suggest that large earthquakes (M > 7.0) can act as external triggers of volcanic unrest, and even eruption. This triggering is attributed to either ground shaking (dynamic stresses) or to permanent ground deformation (associated with static stress changes). However, large earthquakes are rare and testing triggering hypotheses has proven difficult. We use geodetic data to show that the 13 November 2016 Kaikōura earthquake (Mw 7.8) triggered local deformation of up to 11 mm at Taupō volcano, 500 km away, which lasted for approximately twelve days. Using elastic geodetic models, we infer that the observed deformation was caused by either aseismic fault slip or a dike intrusion. We then use strong motion data from the surrounding area to show that the Kaikōura earthquake caused maximum dynamic stress changes in the range of 0.15–0.9 MPa in the vicinity of Taupō volcano and conclude that these dynamic stress changes triggered local faulting or dike activity and the associated deformation at Taupō volcano.