Multidimensional ground reaction forces predicted from a single sacrum-mounted accelerometer via deep learning

There is a gap between the understanding of the mechanisms of sporting injury and the ability to monitor these risk parameters during a game. A single-sensor wearable solution which could be used in lieu of captive laboratory biomechanical instrumentation would be a breakthrough for accurate and valid on-field monitoring of athlete performance and safety. Using archive motion capture and force plate data from The University of Western Australia, with new accelerometer trials conducted at Liverpool John Moores University, this proof of concept exploited deep learning techniques to investigate prediction of 3D ground reaction forces (GRF) from a single sacrum-mounted accelerometer.