Paper Submission & Registration
8th Dutch Bio-Medical Engineering Conference
16:30   Sports
Chair: Chris Baten
15 mins
Measuring trunk motion during on-site wheelchair propulsion using inertial measurement units
Marit P. van Dijk, Marco J.M. Hoozemans, Monique A.M. Berger, DirkJan (H.E.J.) Veeger
Abstract: The interaction between athlete and wheelchair is crucial in all wheelchair sports disciplines. More specifically, trunk posture and movement are reported to affect performance in wheelchair sports and it is suggested that trunk movement is related to power production during wheelchair propulsion [1–3]. However, an ambulatory system to measure changes in trunk inclination in the field is not yet available. Inertial measurement unit (IMU)-based systems seem to be suitable to measure changes in trunk inclination in daily wheelchair sport practice but require complex computations. The aim of this study was to develop a signal analysis method to accurately predict instantaneous trunk inclination during wheelchair propulsion in the field using IMUs. To obtain the three-dimensional orientation from IMU data, a complementary (sensor fusion) filter was used. The use of such a filter has recently shown accurate results in estimating IMU-based trunk orientation [4]. However, the activities used in previous studies involved sports motions without considerable signal distortions. Since signals from IMUs attached to wheelchair athletes are commonly exposed to distortions due to linear accelerations (e.g. every push) and magnetic sources (e.g. material of the wheelchair), applying a complementary filter in a regular way will not meet the required accuracy during wheelchair sports. Therefore, in the current study, a complementary filter was augmented with machine learning to improve its accuracy. Six participants with and six participants without wheelchair experience performed a series of sport-specific field tests with IMUs attached to their wheelchair (frame) and trunk (sternum), while measured simultaneously with a three-dimensional optical motion analysis system as the gold standard. Trunk inclination (represented by the helical angle between the trunk and wheelchair segment) was calculated from the IMU signals and gold standard, which were compared to assess the validity of the IMU-based trunk inclination. Results show root-mean-square errors of less than 6 degrees when IMU-based trunk inclination was compared with the gold standard. In conclusion, a complementary filter augmented with machine learning provides accurate instantaneous trunk inclination estimates during wheelchair propulsion using IMU data. IMUs can thus be validly used to assess trunk inclination during wheelchair sports activities in the field.
15 mins
Elbow muscle activity at the instant of peak valgus load during overhead throwing
Bart van Trigt, Eva Galjee, Marc Hoozemans, Frans van der Helm, Dirkjan Veeger
Abstract: Baseball pitching shows a high prevalence of ulnar collateral ligament injuries, likely due to the high external valgus load on the medial side of the elbow at the instant of maximal shoulder external rotation (MER). Based on in-vitro studies this external valgus load is resisted by the ulnar collateral ligament, but it can be assumed that also muscles contribute to stabilizing the elbow joint. Whether muscles also actively stabilize the elbow during pitching is, however, not yet shown. For this purpose, the aim of this study is to determine whether the lower and upper arm muscles are active at the instant of MER during a fastball pitch. Six recreational uninjured pitchers threw fifteen fastball pitches. Surface electromyography of six elbow muscles were measured at 2000Hz. The pitching motion was obtained with a high speed camera at 240 Hz, to determine the MER. The results show moderate to high activity at MER in all participants for the lexorpronator muscle group and the pronator teres. The extensor-supinator muscle group showed low to moderate activity. The elbow muscles (triceps brachii, biceps brachii and anconeus) show activity at MER in most of the participants, although considerable variation between participants was found, especially in the biceps and triceps brachii muscles. The muscle activation of the flexor-pronator group and pronator teres at MER indicate a direct contribution of forearm muscles in counteracting an external valgus load during pitching. The activation of the biceps and triceps brachii muscles indicates a possible indirect contributory effect as the combined activity of these muscles counteract opening of the humeroulnar joint space. We conclude that active muscular contribution counteracting the elbow valgus load can be presumed and is thus of importance in injury risk assessment. Future research should investigate the load distribution across the elbow stabilizers.
15 mins
Automated monitoring of load exposure and coping in the context of performance in volleyball
Vinish Yogesh, Tom Lankhorst, Chris T.M Baten
Abstract: Introduction: Team sports have become highly competitive and a considerable amount of resources are invested by professional sports companies to analyze the team’s performance and training, e.g. Volleyball is a sport that relies on the biomechanical ability and positional tactics of the players, forging a need for accessing the player kinematics and positional patterns to improve the training and evaluation of the players. Technological advancements such as videography have made these analyses more intuitive. However, current videography solutions involve (subjective and laborious) manual annotation and cannot provide kinematic analysis. Though the advent of movement and position sensors facilitates objective analysis of both player position and biomechanical performance, there are no systems for volleyball assessment. Therefore, this research aims to develop and validate an automated system for the measurement of both the player position and jump parameters in volleyball using one sensor module on the player. Methods: The developed sensor system combined UltraWideBand (UWB) and Inertial Magnetic Measurement Units (IMMU). An Extended Kalman Filter (EKF) was developed for estimation of the player position by sensor fusion of UWB and IMMU data. Jump parameters were calculated by a jump detection algorithm based on the acceleration data and EKF estimates. Results: From preliminary experiments, the position estimates from the EKF showed an average error of 0.28 to 0.30 meters. The RMSE values varied between 0.06 to 0.16 meters for x-axis, 0.21 to 0.27 meters in y-axis and 0.14 to 0.19 meters in z-axis for experiments performed for position validation. The jump detection was 97.5% accurate, while the jump height estimates had a mean error of 0.011m with a standard deviation of 0.019m. Conclusion: The estimates of position were sufficiently accurate for the estimation of player position considering the requirement of a player movement pattern with respect to the different zones in the field and other players, rather than absolute position coordinates. The jump detection and height estimates were highly accurate for the volleyball assessment. Therefore, the proposed and developed system is adequate to function as a potential tool to evaluate volleyball and provides a platform for future volleyball and other team sports assessment.

end %-->