暂无分享,去创建一个
Human-like motion is a primary goal for many robotic assistive devices. Emulating the strategy of the human neuromuscular system may aid the control of such powered devices, yet many challenges remain. In this study, we investigated the potential for using the winding filament model (WFM) of muscle to predict the net muscle moment of the ankle. The long-term goal is to use this model to improve ankle control of a commercial powered exoskeleton. The innovation aspects of this study are: First, there have been no commercialized active ankle exoskeletons available in the market. All the available exoskeletons have passive ankle joints, which cannot mimic human movement, especially in normal and fast walking [1]. Second, the Winding Filament Model Controller (WFMC) is the first control strategy based on a muscle model that does not use an electromyographic (EMG) signal as an input. The activation, which is calculated from EMG, is a crucial input parameter for almost all of the control strategies based on muscle modeling. However, the winding filament muscle model can predict muscle force by using muscle length as the primary input, and the activation input signal could be either a square wave [33] or a simple bell shape function like our study. This is one of the most important benefits of the WFMC strategy, since it makes this bioinspired strategy applicable for all patient populations, even those with impaired muscle activities or without muscle activities (e.g. stroke, spinal cord injury, Parkinson disease, etc.). Third, the WFMC is adaptive to different tasks like walking at different speeds, as well as walking over the incline and over stairs.
[1] D. Pai,et al. Is titin a ‘winding filament’? A new twist on muscle contraction , 2012, Proceedings of the Royal Society B: Biological Sciences.
[2] G. Zahalak. A distribution-moment approximation for kinetic theories of muscular contraction , 1981 .
[3] Jacob Rosen,et al. Performances of Hill-Type and Neural Network Muscle Models - Toward a Myosignal-Based Exoskeleton , 1999, Comput. Biomed. Res..