脑机控制智能轮椅

发布时间：2018-05-11 20:02

本文选题：智能轮椅 + Visual　；参考：《华南理工大学》2016年硕士论文

【摘要】：脑-机接口(Brain Computer Interface,BCI)区别于传统的人机交互方式,给患有运动功能障碍的残疾人或者行动不便的老年人供了一种全新的人机交互方式。随着世界老年化程度的加深,以及受各种自然灾害影响造成的残障人士数量的增加,如今越来越多的研究机构或者高校开始致力于改善这种人群的行动自由度和生活质量,为他们供一种基于脑-机接口的代步工具,即智能轮椅,因此具有自动行走、自主避障、人机交互、地图记忆等功能的智能轮椅越来越受到人们的青睐。本文研究基于共享控制策略的脑控智能轮椅系统,该系统结合BCI控制模式和自主控制模式两种控制模式,将人工智能和机器智能有机的进行融合。自主控制模式下,智能轮椅能通过Laser传感器检测外部环境,通过角度势场法实现自主避障,通过深度RGB视觉传感器获取环境中的视觉特征来辅助定位。采用Vision-based SLAM算法实现轮椅的同步定位和绘图。脑-机控制控制模式,本文突破了传统的基于方向驱动的控制方式,将BCI与轨迹规划算法相结合,实现智能轮椅的平滑的方向调整时沿着曲线运动,避免了轮椅运动过程中的晃动,减少了对脑电信号造成的影响。本文出了基于脑-机接口控制模式和自主控制模式的共享控制策略,通过对比基于模式切换的共享策略和基于状态融合的共享控制策略,分别进行实验测试与分析。本文采用模块化设计思想,将系统的功能模块进行单独的理论仿真分析和实际测试,使得系统的功能模块均具可实施性和稳定性。最后本文征得多名志愿者的同意,对每个志愿者进行了多次测试实验,并且严格按照统一标准统计实验结果、记录实验过程,分析实验结果,最终使得所有志愿者都能够达到本文设计的要求。特别的,为满足以上各系统功能的要求,本文通过对一台普通的电动轮椅进行改装,重新设计了系统的硬件和软件,并且进行了大量的验证实验,具体包括:避障实验、视觉辅助定位实验、脑机信号测试实验、轨迹规划实验和共享控制策略实验等。实验测试表明本文采用的理论算法以及出的方法都具有一定的实际效果,本文设计的智能轮椅具有一定的可行性和实用性。
[Abstract]:The Brain Computer Interface (BCI) is different from the traditional human-computer interaction. It provides a new human-computer interaction for disabled persons with movement dysfunction or inconvenient old people. With the deepening of the world's aging and the increase of the number of disabled people affected by various natural disasters, such as Nowadays, more and more research institutions or colleges and universities are beginning to improve the freedom of action and quality of life of this population, and provide them with an Intelligent Wheelchair Based on the brain machine interface. Therefore, the intelligent wheelchair with automatic walking, autonomous obstacle avoidance, human-computer interaction, map memory and other functions is becoming more and more popular. This paper studies the brain control intelligent wheelchair system based on shared control strategy. This system combines the two control modes of BCI control mode and autonomous control mode to integrate artificial intelligence and machine intelligence. Under the independent control mode, the intelligent wheelchair can detect the external environment through the Laser sensor and realize the obstacle avoidance by the angle potential field method. Using the depth RGB vision sensor to obtain the visual features in the environment to assist the positioning. Using the Vision-based SLAM algorithm to realize the synchronous positioning and drawing of the wheelchair. The brain machine control control mode, this paper breaks through the traditional direction driven control mode and combines the BCI with the trajectory planning method to realize the smooth direction of the intelligent wheelchair. Moving along the curve to avoid the wobble in the wheelchair movement and reduce the impact on the brain signal. This paper presents a shared control strategy based on the brain machine interface control mode and the autonomous control mode. By comparing the sharing strategy based on the mode switching and the state fusion based sharing control strategy, the real time is real. In this paper, the modular design idea is adopted to carry out a separate theoretical simulation analysis and practical test of the functional modules of the system, so that all the functional modules of the system are feasible and stable. The standard statistical experiment results, the record of the experiment process and the analysis of the experimental results, all the volunteers are able to meet the requirements of the design. In particular, in order to meet the requirements of the system functions above, this paper resets a common electric wheelchair, resets the hardware and software of the system, and has carried out a large number of tests. The proof experiment includes: obstacle avoidance experiment, visual assistant positioning experiment, brain machine signal test experiment, trajectory planning experiment and sharing control strategy experiment. The experimental test shows that the theoretical algorithm and the method used in this paper have some practical effect, and the intelligent wheelchair designed in this paper has certain feasibility and practicability.

【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH789

【相似文献】