摘要：人体肌肉骨骼系统的建模仿真，对于仿生式机器人的设计具有重要的指导意义。在本 文中，我们根据人体生物学原理建立了手臂的生物学模型系统并进行了仿真。首先，我们 将人体手臂简化为四个自由度——肩部 3 个、肘部 1 个并运用刚体运动学及其次坐标系 的知识建立了运动学模型。然后，我们分析了人体肌肉的生物学原理，确立了控制手臂的 15 块肌肉及其属性。综合考虑了关节约束力，骨骼的接触力等实际力学环境。综合肌肉 力和环境约束力，基于多刚体动力学的知识，我们建立了手臂的动力学模型。接着，我们 通过 opensim 平台，基于 xml 语言搭建出了模型。最后，我们通过 ESN 网络+增强学习的 方法，基于 VS2013 和 opensim 函数库，用 C++编写出程序，初步实现了对模型的运动控 制，并能实现一种简单的对输出力的控制——抵抗外力。
关键词 肌肉骨骼模型 仿生 刚体 opensim ESN 增强学习
毕 业 设 计 说 明 书 外 文 摘 要
Title Research on Intelligent Motion Control Simulation Platform Construction and Model Based on Bionic Robot
Abstract：The modeling and simulation of the human musculoskeletal system is of great significance to the design of the bionic robot. In this paper, we established the arm's biological model system according to the principles of human biology and carried out the simulation. First, we modeled the arm model into four degrees of freedom - the three shoulders, the elbow. We used the theory of body kinematics and homogeneous coordinates system to establish a kinematic model. Then, we analyzed the biological principles of human muscles, established 15 muscles to control the model and decided their properties. We also considered the joint limit force, the contact force and other physical environment. Based on multi-rigid-body dynamics, we established the dynamic model of the arm. Then, we used the opensim platform to build a model. Finally, we employed on the ESN network + enhanced learning method, based on VS2013 and opensim library, with C + + program, to realize motion control of the model. We could also achieve a simple force control - resistance to external forces.
Keywords musculoskeletal model bionic rigid body opensim ESN reinforcement learning
目 次
1 引言 1
1.1 研究背景及意义 1
1.2.1 仿真平台国内外现状 2
1.2.2 神经控制算法求解国内外现状 2
1.3 章节安排 3
2 运动学模型搭建 5
2.1 手臂自由度分析 5
2.2 模型坐标系建立 6
2.3 模型运动学方程建立 8
2.4 运动学模型检验 10
3 动力学模型搭建 11
3.1 刚体模型的力学属性 基于仿生机器手的智能运动控制仿真平台建设及模型研究:http://www.chuibin.com/jixie/lunwen_206248.html