Manvendra provided helpful links.

It is not too difficult to model an assistive device with OpenSim (though this depends strongly on how detailed you want the model of the device to be). The most important thing to consider is the question you want to answer. Predicting the change in kinematics from wearing a device is very challenging, and usually requires using optimization algorithms to determine the control signals that drive your model to achieve a walking motion (http://journals.plos.org/plosone/art...l.pone.0121407). Two simpler types of problems are:

- (a) estimating the muscle mechanics that occurred during experiments with an assistive device (see these 2 studies: http://jeb.biologists.org/content/ea.../24/jeb.150011, http://biomechatronics.cit.cmu.edu/p...n_2017_JEB.pdf, https://www.ncbi.nlm.nih.gov/pubmed/25278469), and
- (b) predicting device and muscle behavior with *fixed* kinematics (see this paper on running: http://journals.plos.org/plosone/art...l.pone.0163417).

The OpenSim documentation contains two examples that should help you learn what you can do with OpenSim:
- http://simtk-confluence.stanford.edu...Ankle+Injuries
- http://simtk-confluence.stanford.edu...Metabolic+Cost

Can I use the models available in Opensim directly or do I need to collect gait data of a user walking with the same device for simulation in Opensim?
You will be able to use existing OpenSim models of the human lower limbs, but you will need to create your own model of the assistive device.

As mentioned above, whether or not you must collect gait data depends on your question and assumptions (do you assume that kinematics can change?). In general, collecting experimental data (with a device) would be helpful to validate results you obtain from the simulations.

You can find gait data for normal walking on SimTK.org (also, OpenSim comes with sample gait data for walking).

You can also direct your questions to the OpenSim forum: https://simtk.org/plugins/phpBB/inde...uginname=phpBB