一种仿生机械臂空间位置反馈方法的研究

摘要

To solve bionic manipulator space position feedback, electrooculography (EOG) is used as the neurofeedback in the system. The mapping rules between the eye rotation angle and the corresponding EOG are analyzed by the experiment,in which the subjects repeatedly scan the multi-angle targets, and then the location model based on binocular vision of human eye is established; then the targets location information extracted from the parametric location model is converted to the manipulator coordinates via the quaternion coordinate transformation model between human eye and manipulator on the physics experiment platform to guide the positioning. The results show that the location model following binocular vision of human eye enables to effectively extract targets coordinates to provide space position feedback for bionic manipulator; simultaneously it provides a new idea for bionic manipulator space position feedback.%为了解决仿生机械臂空间位置反馈的问题,将眼电图作为神经反馈,应用于仿生机械臂空间位置反馈系统,通过多角度目标点重复扫视实验,分析眼球转动角度与对应眼电信号特征之间的映射规律,建立了人眼双目视觉定位模型.采用四元数法构建人眼、仿生机械臂坐标转换模型,并通过自行设计的物理实验平台进行坐标转换参数标定、计算旋转矩阵和平移向量,根据参数化定位模型将提取到的目标点转换为机械臂坐标系位置信息,从而实现了空间定位.研究结果表明,人眼双目视觉的定位模型可以有效地提取目标点坐标,为仿生机械臂提供空间位置反馈,同时也为仿生机械臂空间位置反馈提供了新思路.