助老助残移动机器人手势交互控制研究

发布时间：2018-06-18 15:42

本文选题：智能移动轮椅 + 手势交互　；参考：《南京邮电大学》2017年硕士论文

【摘要】：随着医疗技术与科研水平的突破,智能助老助残移动服务机器人为行动不便的残障人士提供了出行上的便利。对于老年人尤其是残障人士在使用轮椅时,轮椅的人机交互设计就显得十分之关键。手势交互的实时性、稳定性与易学性深刻地影响到用户的使用体验。本文从这个问题出发设计出一种基于Kinect摄像头的在智能移动轮椅上使用的手势交互控制方法以用来可靠、实时地用手部动作控制轮椅运行。本文的主要研究内容与成果如下:(1)依据手势识别技术的基础理论与方法,在实验平台的硬件结构框架设计与轮椅驱动控制设计背景下,首先对用户的手势交互展开需求分析工作,其次以需求为纲设计出一种虚拟手柄手势交互方法,此方法模拟了实体的操纵连杆,在手部处于不同空间位置时给定相应的行动速度和转角数据,同时虚拟手柄的控制过程使用了图形化界面进行的实时显示。(2)在手势交互控制环节的第一步,应用Kinect摄像头对手势分割与手势跟踪问题进行了研究。在手势分割部分,分析基于高斯肤色模型手势分割方法的不足,设计出基于深度信息与肤色信息分割手势的方法且进行实验。手势跟踪环节使用Camshift算法与Kalman滤波两者结合的跟踪算法,收集视频图像中的手部坐标位置效果良好。(3)在手势交互控制环节的第二步,先进行实现手势空间坐标转换与虚拟手柄机制设定工作;然后对本手势交互控制系统进行模糊控制需求分析、进行模糊控制器设计。完成应用模糊控制对虚拟手柄进行坐标置位控制的工作。(4)搭建智能移动轮椅实验平台,在不同实际场景下进行手势获取干扰性实验和手势控制智能轮椅按指定路线运行的实验。重点分析各类测试结果,检测手势交互控制方法在稳定性和平滑性是否满足设计要求。通过实验证明智能轮椅经模糊置位控制在运行平滑性和按指定路线运行能力上基本达到要求。
[Abstract]:With the breakthrough of medical technology and scientific research, the mobile robot provides mobility convenience for the disabled. When the elderly, especially the disabled, use wheelchairs, the man-machine interaction design of wheelchairs is very important. The real-time, stability and easiness of gesture interaction deeply affect the user's experience. In this paper, a Kinect camera based hand gesture interactive control method for intelligent mobile wheelchair is designed to control the wheelchair in real time and reliably. The main contents and results of this paper are as follows: (1) based on the basic theory and method of gesture recognition technology, under the background of hardware frame design and wheelchair drive control design of the experimental platform, First, the user's gesture interaction is analyzed, and then a virtual handle gesture interaction method is designed based on the requirements. This method simulates the control link of the entity. When the hand is in different space position, the corresponding speed and angle data are given, and the control process of the virtual handle uses the real-time display of the graphical interface. Kinect camera is used to study gesture segmentation and gesture tracking. In the part of hand gesture segmentation, the deficiency of gesture segmentation method based on Gao Si skin color model is analyzed, and the method based on depth information and skin color information is designed and tested. In hand gesture tracking, the tracking algorithm which combines Camshift algorithm and Kalman filter is used to collect the hand coordinate position in video image with good effect. Firstly, the spatial coordinate transformation of gesture and the setting of virtual handle mechanism are carried out, and then the fuzzy control requirement of the gesture interactive control system is analyzed, and the fuzzy controller is designed. The experiment platform of intelligent mobile wheelchair is constructed by using fuzzy control to control the coordinate position of the virtual handle. The experiments of gesture acquisition interference and gesture control intelligent wheelchair running according to the specified route are carried out in different actual scenes. Analysis of all kinds of test results to detect the stability and smoothness of gesture interaction control methods meet the design requirements. It is proved by experiments that the running smoothness and the ability of running according to the specified route can be basically achieved by the fuzzy position control of the intelligent wheelchair.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【参考文献】