基于支持向量机的外骨骼机器人灵敏度放大控制

为了准确控制外骨骼机器人跟随人体运动,需要建立其动态、精确的数学模型；人体下肢外骨骼是一个多自由度、强耦合以及非线性的多连杆系统,难以建立准确的运动学和动力学模型；文章使用三维运动捕捉与空间定位系统,获取实际人体运动参数(运动学与动力学),应用支持向量机(SVM)学习人体下肢外骨骼的数学模型；基于该模型构造基于支持向量机模型的灵敏度放大控制方法；文章使用MATLAB和LIBSVM建立外骨骼下肢机器人的数学模型,并进行仿真分析；仿真结果表明基于SVM的模型学习方法,能够准确计算出人体下肢外骨骼的动力学模型,并简化建模过程；基于SVM的灵敏度放大控制,能够有效计算出人体下肢外骨骼各关节(髋关节、膝关节、踝关节)的输出力矩,并控制外骨骼机器人跟随人体运动。

Support Vector Machine Based Sensitivity Amplification Control for a Lower Extremity Exoskeleton

The human lower extremity exoskeleton is a multiple degrees of freedom, strong coupling and nonlinear multi-link system, it is difficult to establish an accurate kinematic and dynamic models. We use three-dimensional motion capture and spatial positioning system, to get the actual human motion parameters (kinematics and dynamics), use support vector machine (SVM) to learn mathematical model of human lower extremity exoskeleton. Using the model we constructed the control method of support vector machine based sensitivity amplification. Using MATLAB and LIBSVM to build the model, simulation results show that the learning method based on SVM model and SVM based sensitivity amplification control, can accurately calculate the dynamic model of the human lower limb exoskeleton, and control the exoskeleton robot to follow the movement of the human body.