-
摘要：在科技不断进步的今天，有越来越多的机器人被应用到各个领域，对于移动机器人等作为较为基础的机器人的研究日趋完善。而移动机器人的故障诊断也成为机器人研究领域的一个重要的部分。大部分对于移动机器人的工作状态研究都是建立在整个系统能够以理想状态工作的基础上。但是移动机器人在真实的环境中工作，面对多变的周遭环境，发生故障的概率还是比较大。特别是在复杂的未知环境中，移动机器人更加容易发生故障，而此时我们也无法或很难对其进行文护文修性的干预活动。如若故障没有检测出来或者没有得到及时的处理，移动机器人将会完全陷入一个未知的高危工作状态中。这不仅会严重影响移动机器人的使用寿命，更有可能会使移动机器人不能进行正常活动，甚至可能会导致灾难性的后果。
- 上一篇：基于DSP的异步电动机直接转矩控制系统设计
- 下一篇：MATLAB/SIMULINK不稳定过程的智能控制及其应用仿真
本论文针对如今移动机器人常见的故障，提出较有效可行的诊断方法，并对其进行研究。本文概述了移动机器人故障诊断技术的研究现状。详细讨论了基于神经网络的移动机器人的故障诊断问题。神经网络能够类似人脑一样，通过自适应控制器能够在线或离线学习不断地改进系统的性能，提高神经网络的诊断精度，从而达到最优的故障诊断。并且整个控制过程可以不需要系统的数学模型知识。同时，神经网络的并行结构又可以帮助解决很高的计算要求。
本文根据各模块传感器之间的数据信息，归纳了一个故障分类方法，将移动机器人的故障基本分为系统故障、传感器故障和混合故障三类。并在此分类方法提出了一个基于RBF神经网络的故障诊断处理方法。该故障诊断方法的基本思路是将经过处理后的各模块组的传感器信息作为神经网络的输入，故障类型作为输出，利用神经网络完成各模块组的故障诊断过程。而利用故障类型的输出进行的仿真实验有效地证明了该故障诊断方法在移动机器人故障诊断上的可行性。
最后，对本文所提到的研究进行总结，同时也对移动机器人故障诊断方法的拓展作了展望。
关键词：移动机器人 ；故障诊断 ；神经网络
The Research of Fault Diagnosis Based On Neural Network For Mobile Robot
Abstract: In today’s technological advancements, a growing number of robots is being applied to various fields, while the research on basic robots, such as mobile manipulator robots, is improving. Fault diagnosis for mobile robot has also become an important part of the field of robotics. Most of the mobile robot research was based on work in the system as a whole to the ideal State. Mobile robot working in a real environment, in the face of changing surroundings, the probability of failure is quite large. Especially in complex in an unknown environment, mobile manipulator more failure-prone, but we cannot at this time or it's hard to make maintenance and repair interventions. If failure is not detected or does not receive timely treatment, move the robot into a high-risk status of unknown. This will not only seriously affect the life of mobile manipulators, are more likely to have mobile manipulators that do not engage in normal activities, and might even lead to disastrous consequences.
Common failure for this article for now, mobile manipulator, more effective diagnostic methods, and study it. This article provides an overview of fault diagnosis for mobile robot research. A detailed discussion on fault diagnosis for mobile robot is based on neural network. Neural network similar to the human brain, through adaptive controller to online or offline learning and constantly improving the system's performance, improve the diagnostic accuracy of the neural network, so as to achieve optimal fault diagnosis. And the whole control process does not require knowledge of mathematical model of the system. At the same time, parallel neural network structure can help to address very high computational requirements.