Towards “intelligent“ loose garment for detecting body posture in rehab

Project start: 

01 June 2012

iCareNet Fellow: 

Abstract of the Project

Within the rehabilitation of stroke patients, several assessment questionnaires exist that allow to measure the performance of the patients. These are based on rating the execution of everyday tasks in a controlled environment by a therapist which is quite subjective. However, current activity recognition research to measure this performance objectively is mostly relying on body worn sensors that are quite obtrusive.

An unobtrusive measurement of the body posture with textile-integrated sensors will be the focus of this project.

Design and development of flexible MEMS acceleration sensors

With the help of dedicated tools (e.g. COMSOL), different sensor designs for flexible MEMS accelerometers will be simulated. The sensor sensitivity and the effects of bending will be identified. Using ETH resources, i.e. the clean room, a process to produce the sensors will be developed. These sensors will then be characterized and potential for optimization identified.

Integration of sensors into textiles

To integrate the sensors into the textiles, a process to structure heat shrink films with conductive paths will be established and its parameters identified. The sensors and related evaluation electronics will be connected on to these films which are then applied to the textile. In a second approach the sensors shall be integrated on flexible plastic stripes that are then woven into the textile.







Fig 1: Design of 1D accelerometer [1]

Estimation of body posture

Building upon existing models on drape of loose garment, these will be modified and extended to model the behaviour of the textile with the sensors. By using multiple sensors at one body location, an algorithm will be developed to try to compensate for the sensor orientation error in loose garment. These results will finally be verified in a patient study together with Reha Rheinfelden.







Fig 2: Sensor orientation error [2]

Research Achievements to date:

  • Literature review and refinement of research topic



[1] Gönenli et al., Surface Micromachined MEMS Accelerometers on Flexible Polyimide Substrate, IEEE Sensors Journal, vol. 11, no. 10, 2011

[2] Harms et al., Estimating Posture-Recognition Performance in Sensing Garments Using Geometric Wrinkle Modeling, IEEE Transactions on Information Technology in Biomedicine, vol. 14, no. 6, 2010