Starting date: 2019
Duration: 20 months

Project category

OPTION 1: develop a benchmarking solution for one specific benchmarking scenario including one or more outcomes: testbed, software routines and/or experimental datasets.


Wearable robots


Testbed, Sofware & Dataset



  • Delft University of Technology (The Netherlands)

Partner 2:

  • University Rehabilitation Institute (Slovenia)

Partner 3:

  • GTXmedical (The Netherlands)
To quantify balance of bipedal gait, for example to benchmark exoskeletons, prostheses or bipedal robots, a promising approach is to induce perturbations and to register the resulting subject responses. However, protocols across research groups and perturbation devices are hardly unified. Between the project partners, there are two very different experimental platforms: the wearable GyBAR which can induce torques to the upper body, and the BAR-TM, a treadmill-based device that can apply lateral forces to the hip. These devices enable to test a wide range of perturbation strategies and measure the reactions. We want to investigate whether assessment with similar precision and accuracy can be performed with simpler devices while also investigating unified protocols to warrant the comparability of results across platforms. The unified protocols are expected to be independent of the characteristics of a perturbation device and should therefore be easier to reproduce. The specific objectives of the proposed work are:
a) develop experimental protocols and outcome measures to assess the ability to recover from perturbations and balance, using these two devices,
b) to distil a simpler version consisting of wearable sensors and off-the-shelf perturbation solutions, for example, using visual or self-induced perturbations, to allow widespread use of such protocols,
c) crossvalidate perturbation assessment protocols between the sophisticated devices and the simpler version using appropriately collected datasets.
We intend to make the protocols and hardware available for benchmarking the balance performance of a range of bipedal systems across multiple scenarios and hence wish to integrate it with the EUROBENCH framework, along with the necessary software routines and appropriate datasets.