Measuring human movement is important with regard to identifying inappropriate techniques which may predispose to injury eg. running technique. Optical motion capture systems are usually seen as a gold standard for accuracy in monitoring human movement. These systems are typically very costly and fixed in a laboratory environment. On the other hand, wearable inertial sensors are a low cost, more accessible alternative for laboratory free motion capture and analysis. Limitations to accuracy still exist as inertial sensors suffer from noise and biases which lead to errors in accuracy which accumulate over time.
Clare’s research aims to design a novel algorithm to address existing limitations to the accuracy of inertial sensor data for human motion capture, with a focus on calculating accurate data for a running based application. This involves the added complication of accuracy in the presence of high impact accelerations which occur during running, as well as sensor bias affecting the accuracy of the calculated movement data. Additionally, it is required that data is accurate over a long period of time e.g. an hour run, for better analysis of the movement. An algorithm which gives real-time results would be very beneficial as it has been shown that near real-time feedback can help running technique retraining. Clare has developed two android applications, SoftRun and SoftLand, to show how real-time feedback using wearable inertial sensors can help reduce impact loading in both landing (SoftLand) and foot strikes (SoftRun) while running.