Using detailed musculoskeletal models in computer simulations of human movement can provide insights into individual muscle and joint loading; however, these muscle models increase problem dimensionality and require difficult-to-fit parameters. Here, we provide a brief overview of a muscle model alternative, muscle torque generators (MTGs), and highlight how MTG functions have been used by researchers to generate accurate dynamic simulations of optimal sports performance. Multibody dynamic models of a golf drive, track cycling, and wheelchair propulsion were designed and actuated using MTGs. Each MTG was effectively a rotational, single muscle equivalent that contained joint angle/velocity scaling and passive elements to mimic Hill-type muscle model behaviour. Optimal control algorithms were used to predict how each model would execute their respective sports task; these results were compared against experimental data collected from elite athletes. Good agreement between simulated and experimental movement trajectories was observed, with relatively low computational times required for convergence of the MTG-driven multibody simulations.

在人体运动的计算机模拟中使用详细的肌肉骨骼模型可以洞悉单个肌肉和关节的负荷；但是，这些肌肉模型增加了问题的维度，并需要难以拟合的参数。在这里，我们简要介绍了肌肉模型替代产品，肌肉扭矩生成器（MTG），并着重介绍了研究人员如何使用MTG功能来生成最佳运动表现的准确动态模拟。使用MTG设计并启动了高尔夫，赛道骑行和轮椅推进的多体动力学模型。每个MTG实际上都是一个旋转的单肌肉等效物，其中包含关节角度/速度缩放和被动元素，以模仿Hill型肌肉模型的行为。最优控制算法被用来预测每个模型如何执行各自的运动任务。将这些结果与从精英运动员那里收集的实验数据进行了比较。观察到模拟运动轨迹与实验运动轨迹之间的一致性好，而MTG驱动的多体模拟的收敛所需的计算时间相对较短。