Estimation of joint contact forces using a real-time forward dynamics EMG-driven musculoskeletal model
Date
2012
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Gait research is an important part of gaining a better understanding of the onset and progression of many debilitating diseases. There are many different tools used in this field, but one of significance is forward dynamics modeling. This technique has allowed researchers to gain insight into the force generation of different muscles during gait. A shortcoming to this modeling technique is that all data must be collected before it can be analyzed and processed offline. This provides no significant returns until several hours after the data has been collected. In academic gait research this is not a problem, but more often than ever we are seeing an effort to create a more synergistic relationship between the clinic and the laboratory setting. Researchers are working hand in hand with clinicians to help bring their findings into immediate practice. It is for this reason that a real-time paradigm is the next logical application of such modeling practices. Real-time modeling of joint forces would allow clinicians to step from a largely qualitative rehabilitation regime, to a more quantitative understanding of patient specific kinetics thus allowing them to increase their therapeutic efficacy. The goal of this project is to implement a model which will calculate the forces that the muscles exert on the joint, and the joint reaction force from inverse dynamics in real-time. From these two components it is possible to calculate the total joint contact force experienced by the tibial plateau. The estimation of muscles forces will be done through the use of an existing forward dynamics EMG-driven musculoskeletal model. The model used for real-time force estimation is a derivative of the model presented by Buchanan et al. (2004) that, in order to operate in real-time, does not take the force-velocity component of muscle contraction characteristics into account. For this reason an application for this real-time approach has been presented that looks to minimize the force-velocity component and utilize the real-time estimation of joint contact force. The protocol presented, standing target matching (STM), has previously been used to show differences in muscle co-contraction in subject populations where activation patterns have been suggested to differ from healthy individuals: ACL deficient and osteoarthritic patient populations. The STM protocol presented aims to look at varying levels of joint contact force experienced at different target locations across a small subject population of three people as a proof of concept. The results have been presented and the implementation of the model has been discussed. Overall the ability to estimate total joint contact in real-time is a powerful tool and could be applied to a large array applications.