Search

We aim to demystify the development of neuro-biomechanical modeling in OpenSim with zero configuration, easy to share models while accessing to free GPUs on a web-based platform of Google Colaboratory. OpenSim is an open-source biomechanical package. OpenSim is used in a variety of applications and developed in C++; however, it is available for a wide range of users with bindings in MATLAB, Python, Jython and Java via OpenSim Application Programing Interface (API). OpenSim installation on a personal computer is well described by the developers but its implementation may still be time-consuming and challenging for the new users. Cloud-based computing is expanding in almost all engineering domains with zero configuration, though it is in its early stages within biomechanics community. In this study, we aim to access OpenSim functionality on the Google cloud platform. The methods can also be used in other cloud-based platforms. We installed OpenSim on the Google Colab via Anaconda cloud and named it OpenColab. To use OpenColab, one requires only a connection to the internet and a Gmail account. Moreover, such a platform enables the users to access vast libraries of machine learning available within free Google products e.g., TensorFlow. OpenColab takes advantage of zero configuration of cloud-based platforms even on a smart phone, provides access to free GPUs and enables users to share and reproduce modeling approaches for further validation. Finally, we performed inverse problem in biomechanics and compared OpenColab results with OpenSim GUI’s for validation. Step-by-step installation processes and examples can be found freely at: https://simtk.org/projects/opencolab.