外骨骼机器人体感交互控制与仿真  

Simulation and interactive somatosensory control for exoskeletal robot

作　　者：瞿畅[1] 于陈陈 沈芳[1] 

机构地区：[1]南通大学机械工程学院,南通226019

出　　处：《现代制造工程》2017年第7期42-46,共5页Modern Manufacturing Engineering

基　　金：南通市前沿与关键技术创新计划项目(MS22015007);南通大学研究生科技创新项目(YKC14007)

摘　　要：为使外骨骼机器人应用于康复训练时,其动作控制更便捷且符合人体运动习惯,建立了一种通过Kinect体感设备实现外骨骼机器人体感交互的运动控制方法。通过Kinect获取人体骨架关节点,构建空间向量,依据人体坐标系对上肢肩、肘关节运动角度进行分解计算,获取体感控制角度。采用卡尔曼滤波算法对角度数据进行平滑处理,改善角度数据的稳定性和实时性,并通过限速和限幅处理,确保运动的安全性。最后,将角度数据组传递到基于Lab VIEW建立的外骨骼模型,进行实时控制,根据外骨骼模型的运动效果,验证了数据计算与处理的正确性和可靠性,为建造物理样机奠定基础。A new method was proposed on controlling exoskeletal robot by using Kinect sensor in order to make the movement of exoskeletai robot more conforms to human motion. With vector constructed by skeletal data, the angle for somatosensory control was acquired through dividing and calculating the space angles of upper limb joint according to the coordinate system of human body. Kalman filter was used to smooth the angles and the result showed that Kalman filter can better improve the stability and re- al time property of data. The maximum of angular velocity and the ranges of angles were set up for security. Lastly ,the processed data was passed to the exoskeletal robot model which was built up based on LabVIEW for real-time control. According to the effect of real-time control ,the data processing method was accurate and reliable that provides reliable references for the development of the physical prototype.

关 键 词：外骨骼 体感交互 平滑处理 实时控制 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]