Development, evaluation and application of a novel markerless motion analysis system to understand push-start technique in elite skeleton athletes

This study describes the development, evaluation and application of a computer vision and deep learning system capable of capturing sprinting and skeleton push start step characteristics and mass centre velocities (sled and athlete). Movement data were captured concurrently by a marker-based motion capture system and a custom markerless system. High levels of agreement were found between systems, particularly for spatial based variables (step length error 0.001 ± 0.012 m) while errors for temporal variables (ground contact time and flight time) were on average within ± 1.5 frames of the criterion measures. Comparisons of sprinting and pushing revealed decreased mass centre velocities as a result of pushing the sled but step characteristics were comparable to sprinting when aligned as a function of step velocity. There were large asymmetries between the inside and outside leg during pushing (e.g. 0.22 m mean step length asymmetry) which were not present during sprinting (0.01 m step length asymmetry). The observed asymmetries suggested that force production capabilities during ground contact were compromised for the outside leg. The computer vision based methods tested in this research provide a viable alternative to marker-based motion capture systems. Furthermore, they can be deployed into challenging, real world environments to non-invasively capture data where traditional approaches are infeasible.

Extended dynamic stiffness model for analyzing flexure-hinge mechanisms with lumped compliance

This paper introduces an extended dynamic stiffness modeling approach for concurrent kinetostatic and dynamic analyses of planar flexure-hinge mechanisms with lumped compliance. Firstly, two novel dynamic stiffness matrices are derived for a flexure hinge connected to rigid bodies by shifting its end node to the mass centre of rigid bodies considering the geometric effect of rigid motion. A straightforward modeling procedure is then proposed for the whole compliant mechanism by selecting the displacements at both the mass centre of rigid bodies and the rest end nodes of flexure hinges as the hybrid state variables, differing from the traditional finite element method. With the presented approach, the statics and dynamics of flexure-hinge mechanisms with irregular-shaped rigid body in complex serial-parallel configurations can be analyzed in a concise form. At last, the presented method is compared with other theoretical models, finite element simulation and experiments for three case studies of a bridge-type compliant mechanism, a leveraged XY precision positioning stage and a Scott-Russell-mechanism-based XY𝑣 flexure manipulator. The results reveal the easy operation and satisfying prediction accuracy of the presented method.

Can Markerless Pose Estimation Algorithms Estimate 3D Mass Centre Positions and Velocities during Linear Sprinting Activities?

The ability to accurately and non-invasively measure 3D mass centre positions and their derivatives can provide rich insight into the physical demands of sports training and competition. This study examines a method for non-invasively measuring mass centre velocities using markerless human pose estimation and Kalman smoothing. Marker (Qualysis) and markerless (OpenPose) motion capture data were captured synchronously for sprinting and skeleton push starts. Mass centre positions and velocities derived from raw markerless pose estimation data contained large errors for both sprinting and skeleton pushing (mean ± SD = 0.127 ± 0.943 and −0.197 ± 1.549 m·s−1, respectively). Signal processing methods such as Kalman smoothing substantially reduced the mean error (±SD) in horizontal mass centre velocities (0.041 ± 0.257 m·s−1) during sprinting but the precision remained poor. Applying pose estimation to activities which exhibit unusual body poses (e.g., skeleton pushing) appears to elicit more erroneous results due to poor performance of the pose estimation algorithm. Researchers and practitioners should apply these methods with caution to activities beyond sprinting as pose estimation algorithms may not generalise well to the activity of interest. Retraining the model using activity specific data to produce more specialised networks is therefore recommended.

The rotary component of leg force during walking and running

The component of ground reaction force (GRF) acting perpendicular to the leg in the sagittal plane during human locomotion (acting in a rotary direction) has not been systematically investigated and is not well understood. In this paper, we investigate this rotary component of the GRF of 11 human subjects (mean age ± s.d.: 26.6 ± 2.9 years) while walking and speed walking on a treadmill, along with eight human subjects (mean age ± s.d.: 26.3 ± 3.1) running on a treadmill. The GRF on both legs was measured, along with estimates of the subject's mass centre and the centre of pressure of each foot to yield total leg lengths and leg angle. Across all steady walking and running speeds, we find that the rotary component of the GRF has significant magnitude (peak values from 5% to 38% of body weight, from slow walking to moderate running, respectively) and implies leg propulsion of the mass centre in the rotary direction. Furthermore, peak rotary force magnitude over stance increases with locomotion speed for both walking and running ( p < 0.05), and the time-averaged (mean) rotary force shows a slight increase with walking speed (though the mean force trend is uncertain for running). Also, an estimate of average power input from the rotary force of the leg acting at the mass centre shows moderate and strong positive correlation with locomotion speed for running and walking respectively ( p < 0.05). This study also shows that the rotary force acts differently in walking versus running: rotary force is predominantly positive during running, but during walking it exhibits both positive and negative phases with net positive force found over the whole stride.

“Mass Centre” Vectorization Algorithm for Vehicle’s Counting Portable Video System

Vehicle counting is one of the most basic challenges during the development and establishment of Intelligent Transport Systems (ITS). The main reason for vehicle counting is the necessity of monitoring and maintaining the transport infrastructure, preventing different kind of faults such as traffic jams. The main applied solution to this problem is video surveillance, which is presented by different kind of systems. Some of these systems use a network of static traffic cameras, expensive for establish and maintain, or mobile units, fast for redeployment, but fewer in diversity. In this paper, one particular concept of a low-cost mobile vehicle counting system is investigated, which uses an object detection method based on calculating “mass centre” of detected features of possible object. A hypothesis of improvement of the basic algorithm was formulated and a modification was proposed. In order to prove the hypothesis, both basic and modified algorithms were tested and evaluated.

Study about the Vehicle Orientation on the Trajectory during Transient Movements

In this paper, it is deduced the ordinary differential equation that results in angle variation between the vehicle mass centre rate and the vehicle fore-aft axis. There are presented the wheels turning angles variations dependent on the angle that gives the vehicle orientation on trajectory. There are studied the following cases: the transfer from linear to circular motion, the transfer from circular to linear motion and the movement amidst poles. For each of these motions, there are presented the vehicle orientation angle variations on trajectory for three distinct rates. It is also presented the travelling rate influence on how the vehicle is situated on trajectory.