Bend sensors modeling for fast signal recovering in human motion analysis

Investigation on the more suitable technologies to register human body movements in 3D space with great spatial accuracy is a very challenging task, because a wide range of applications are concerned, from registration of post-stroke rehabilitation or sports performance, to monitoring of movement of disabled or elderly people, only to give some examples. In this paper, the possibilities offered by resistive bend sensors to register joint bend angles, for human posture recognition and motion analysis, have been explored. This paper suggests to extract an electrical model either to simulate their electrical behavior during bending and extension movements, and to recover the true bending angles from model simulation. To give an example, the proposed model was applied to track human joint movements, and it was demonstrated that it can lead to recover the original signal waveform, which represents the true joint rotation, also for the fastest human speed.