Stride information monitoring and sensing in sports

摘要

Accurate measurements of athletes'' stride parameters, such as stride/step length, stride/step frequency, stance times and foot contact times are important to coaching support and biomechanics research in sprinting. Existing stride parameter monitoring approaches are either expensive, or insufficiently accurate, or not suitable for supporting daily training sessions. This paper investigates the use of cost-effective, low quality track-side and on-body sensors in a novel way to enable practical capture of accurate stride information of sprinters. It was investigated how information fusion could play a big part in practical sensing systems; and how novel system design could avoid the need of sophisticated synchronisation between intra-and inter-homogeneous and heterogeneous subsystems. The work also explored the importance of linking heterogeneous data streams to support easy information fusion. Experiment results show that the presented system has an accuracy of 1.661± 1.002cm (mean ± SD) for stride length measurements, with 50% errors fall within 3ms and 80% errors fall within 5ms for foot contact time.