Be like a pro. Play like a pro. Developed by scientists at the Free University Berlin, this sensor uses machine learning and artificial intelligence to learn and train any movement - such as the perfect golf swing. The Free University Berlin is looking for commercialization partners.
1. To see the amended claims filed after receipt of (European) search report go to: https://register.epo.org/application?number=EP15767429&lng=en&...
The invention regards a method and a feedback system for the assessment of motions of a versatile system that use a wireless sensor network attached t...
The invention regards a method and a feedback system for the assessment of motions of a versatile system that use a wireless sensor network attached to the versatile system. The method comprises a training phase and a feedback phase. In the training phase, a defined motion is conducted by the versatile system, whereby sensor data are gathered by the sensor nodes (31, 32, 33). After segmentation (202) of the sensor data into separate segments and the extracting (203) of features, a symbol is assigned to each of the motion fragments. A symbol sequence is formed which represents the complete motion conducted by the versatile system at a sensor node. After repeating at least once the conducting step and/or the analysing step, a reference symbol sequence (207) is formed at each sensor node (31, 32, 33). In the feedback phase, the defined motion is conducted again, and the collected sensor data are analyzed in a similar manner as in the training phase, wherein at each sensor node at least one motion fragment of the present motion is assessed (404) at least by comparing the symbol assigned to the motion fragment with a to be expected symbol of the reference symbol sequence. A feedback (405, 409, 410) is provided to the versatile system depending on the result of the assessment. The feedback refers to a motion fragment and may be provided while the motion is still carried out.