Study On Virtual Rehabilitation System Based On Multi-modal Interaction And Feedback

Generally, patients suffering from the motor function disorders are always tired of the rehabilitation process which is long and tedious. Besides, the heavy economic and psycho logical pressure of the patients also generated from the high cost of the traditional method which is restricted by the space and equipment. With the development of the virtual reality technology, how to apply the technology into the field of rehabilitation and provide a kind of virtual rehabilitation system which is applicable to home and inexpensive has become the main problem in the field of rehabilitation. In the consideration of the immersion, interaction and feedback of the system, the article introduces the following several aspects:Firstly, a low-cost virtual rehabilitation system is designed with completing a variety of hardware communication, data analysis, human-computer interaction and visual design. While the rehabilitation completed through interactive games with the action scenes the patient can get more pleasure and obtain a wealth of physiological feed-back information at the same time.Secondly, the recognition algorithm for complex rehabilitation actions is proposed. th e time series similarity measure with the property of scale, translation and rotation invaria nt has been realized by the dual- threshold value criterion linear approximation pretreatmen t based on the limitation of the length, spatial alternation conversion, and discrete time-seri es graph matching algorithm that follows the time sequence. It improved the existing simp le rehabilitation action recognition algorithms, the recognition accuracy and the interactive performance. Meanwhile, it also provides a variety of rehabilitation motion and levels of difficulty for the patients to choose.Finally, the algorithm for motion smoothing measure based on the curvature estimatio n is presented. Using the product of the local structure of the velocity curve covariance ma trix characteristic root to estimate the curvature and then achieve the quantitative measure for the smoothness of movement based on the estimated curvature. Different situations of r ehabilitation actions were discussed, providing reference to the patients in rehabilitation training to complete the evaluation of the quality of the action, in order to facilitate timely a djustment training status.