Interaction Between Biology and Robotics

In this project we study a salient feature of human motor skill learning that is the ability to exploit similarities across related tasks. We test this concept in forward dynamics simulations of multiple reaching movements of a musculoskeletal arm model with eleven muscles. The eleven muscle activation patterns are generated by a combination of parametrized muscle synergies, where these synergies encode a shared prior among multiple reaching directions. <br/><br/>This project includes the following software/data packages: <br/> <ul> <li> <a href="https://simtk.org/frs?group_id=898#pack_1494">FwdDynSimArmModel2013 </a> : Project files: osim model file, Matlab mex-function and Matlab test file. The musculoskeletal model is based on the upper extremety model by Holzbaur. Geometry files can be found in related simtk projects (e.g. search for Dynamic Upper Extremity Model). </li> </ul>

在本项目中，我们探讨人类运动技能学习的一个显著特征，即能够利用相关任务之间的相似性。通过在具有十一块肌肉的肌骨骼手臂模型的多次伸手运动的前向动力学模拟中测试这一概念，这些肌肉的激活模式由参数化的肌肉协同作用组合生成，其中这些协同作用编码了多个伸手方向之间的共享先验。本项目包含以下软件/数据包：
<ul>
<li><a href="https://simtk.org/frs?group_id=898#pack_1494">FwdDynSimArmModel2013</a>：项目文件：osim模型文件、Matlab mex函数和Matlab测试文件。该肌骨骼模型基于Holzbaur的上肢模型。几何文件可在相关的simtk项目中找到（例如，搜索动态上肢模型）。</li>
</ul>