Dynamic control of a moving platform using the CAREN system to optimize walking in virtual reality environments.

Virtual reality (VR) technology offers the opportunity to expose patients to complex physical environments without physical danger and thus provides a wide range of opportunities for locomotor training or the study of human postural and walking behavior. A VR-based locomotor training system has been developed for gait rehabilitation post-stroke. A clinical study has shown that persons after stroke are able to adapt and benefit from this novel system wherein they walk into virtual environments (VEs) on a self-paced treadmill mounted on a platform with 6 degrees of freedom. This platform is programmed to mimic changes in the terrain encountered in the VEs. While engaging in these VEs, excessive trunk movements and speed alterations have been observed, especially during the pitch perturbations accompanying uphill or downhill terrain changes. An in-depth study of the subject's behavior in relation to the platform movements revealed that the platform rotational axes need to be modified, as previously shown by Barton et al, and in addition did not consider the subject's position on the treadmill. The aim of this study was to determine an optimal solution to simulate walking in real life when engaging in VEs.