A multimodal approach for insole motion measurement and analysis

AbstractIn order to perform long-term, low-effort gait analysis an instrumented shoe insole, equipped with an embedded data processing system, a variety of sensors and a wireless data transmission module has been developed. By using specially developed signal fusion algorithms, the sensor’s raw data, like pressure, 3D tilt and acceleration, is processed to provide information about the user’s gait or moving behaviour. This shoe insole will also form the basis of an “easy to use” balance training system for older people, in order to help decreasing their risk of falling.A set of early prototypes, in the form of two pairs of self-designed and crafted instrumented shoe insoles, has already been developed. For the validation of their functionality a small series of tests, with five users, already took place. A specific test battery was created, consisting of ten tasks, where the stability of gait and coordinative skills were observed. The tasks were compiled from state-of-the-art mobility test and fall assessment strategies. The tests should, on the one hand, prove the stability of the hardware (to endure such testing), the reliability of the wireless connection and should at the same time show, if the procedure of walking can be reproduced from the gathered data. On the other hand these tests were used for gathering gait data from different subjects, which will be used in the future for the training of classification algorithms.Furthermore there is the opportunity that such a wearable system can be used for sophisticated running analysis with qualitative parameters. By logging and/or wirelessly transmitting information about pressure distribution, the course of e.g. the COP and the pitch angle as well as quantitative parameters like time of action, cadence and the number of steps e.g. an activity protocol can be created