机器人仿生双眼运动控制模型研究

发布时间：2018-01-12 08:06

本文关键词：机器人仿生双眼运动控制模型研究　出处：《山东大学》2017年硕士论文　论文类型：学位论文

【摘要】：眼睛作为人类感知和获取外界信息的重要器官,在人类生存进程当中具有举足轻重的作用。机器视觉的出现使得机器智能设备能够具有与生物相似的视觉能力。在此基础上,如何使机器人的双眼能够与人眼一样具有灵活运动、准确定位视觉目标以及快速变化视觉运动模式成为了人们研究的热点。而人类视觉系统全面的功能特点和运动机理给了我们很好的启发。本文从神经生理学和解剖学的角度分析了人类视觉系统结构特性以及分析了不同运动模式的神经通路,重点进行仿生眼运动控制算法研究实现。首先,简单介绍了关于机器人仿生眼研究的背景及意义,对近年来相关的研究成果进行回顾和分析,指出了现有研究存在的问题并对本文主要的研究情况进行了阐述。分析了人类眼球解剖结构和视觉形成的传导通路,从神经生理学层面获取灵感对眼球运动的各种模式的特点和形式进行了总结,分析了相关运动的神经机制及信号传递通路,讨论了仿生眼运动控制与视觉感知的关系。本文以眼球各运动模式的神经生理学结构为基础,分析各神经模块间的作用关系,提炼各模块的数学表达以进行工程模拟。搭建扫视运动模型,通过仿真分析各个模块的作用效果,并与生理学结果进行对比验证;考虑注视点水平方向转移过程中可能存在视觉深度位置变化的情况,从而建立扫视运动与异向运动相结合的复合运动控制模型;考虑眼球平滑追踪运动具有精确注视的特性,建立自适应平滑追踪控制模型,将视网膜滑动误差约束在一定范围之内,提高了跟踪的精确性,能够对匀速、匀加速和随机运动的目标进行准确追踪,解决了现有平滑追踪控制算法仅能对周期性运动进行准确跟踪的问题;在现有头眼协调运动控制模型的基础上进行改进,考虑反射性眼球运动在其中的作用,同时获取头部位置信息与头部速度信息使其易于工程实现。
[Abstract]:The eye is an important organ for human to perceive and obtain external information. Machine vision plays an important role in the process of human survival. The appearance of machine vision enables the machine intelligent equipment to have visual ability similar to that of biology. On this basis. How to make the robot's eyes as flexible as the human eyes. Accurate location of visual target and rapid change of visual movement mode have become the focus of research, and the comprehensive functional characteristics and motion mechanism of human visual system have given us good inspiration. From the perspective of anatomy, the structural characteristics of human visual system and the neural pathways of different motion patterns are analyzed. Focus on the implementation of bionic eye motion control algorithm. Firstly, the background and significance of the robot bionic eye research are briefly introduced, and the related research results in recent years are reviewed and analyzed. This paper points out the problems existing in the present research and expounds the main research situation in this paper. The anatomical structure of human eyeball and the conduction pathway of visual formation are analyzed. The characteristics and forms of various patterns of eye movement were summarized from the aspect of neurophysiology, and the neural mechanism and signal transduction pathway of related movements were analyzed. This paper discusses the relationship between bionic eye motion control and visual perception. Based on the neurophysiological structure of each movement pattern of the eyeball, this paper analyzes the relationship between the neural modules. Abstract the mathematical expression of each module to carry on the engineering simulation, build the scanning motion model, analyze the function effect of each module through the simulation, and compare with the physiological result to verify; Considering that the position of visual depth may change in the horizontal direction of fixation point, a composite motion control model is established. Considering that the eye smooth tracking has the characteristics of accurate gaze, an adaptive smoothing tracking control model is established, which restricts the retinal sliding error within a certain range, improves the accuracy of tracking, and can achieve uniform speed. The target with uniform acceleration and random motion is tracked accurately, which solves the problem that the existing smoothing tracking control algorithm can only track the periodic motion accurately. Based on the existing head-eye coordinated motion control model, the role of reflex eye movement is considered, and the head position information and head velocity information are obtained so that it is easy to be realized in engineering.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【相似文献】