基于神经网络PID的挖掘机轨迹控制系统的实验研究
发布时间：2018-11-05 10:53
【摘要】：挖掘机以其高性能、高效率等优点在建筑领域得到了广泛地应用,但是在目前挖掘机作业过程中,绝大多数依靠驾驶员手动操作,且不宜在危险环境或者长时间工作,实现挖掘机的自动化与智能化显得尤为重要。本课题以实验室挖掘机实验平台为研究对象,旨在通过设计一种高性能的控制器来实现挖掘机工作装置轨迹的高性能控制。本文研究内容主要包括挖掘机工装轨迹控制系统的建模,控制器的设计仿真以及实验研究。在挖掘机控制系统建模过程中,对于动臂和斗杆控制系统分别进行建模,计算各环节的传递函数。液压系统动力机构环节对于液压缸活塞正反向运动分别进行建模,机械结构环节采用最小二乘法拟合求取其传递函数。为实现数字计算机控制,对连续传递函数进行离散化,求取其脉冲传递函数和差分方程。基于建立的挖掘机工装轨迹控制系统的离散数学模型和增量式PID控制算法,设计基于普通PID控制器的变速积分数字PID控制器和由神经网络实现控制器参数调节的单神经元自适应PID控制器、BP神经网络PID控制器,并在MATLAB软件中编写控制器的控制算法程序,完成控制器的参数整定及仿真,得到动臂控制系统和斗杆控制系统在各个控制器作用下,对于测试信号的响应曲线及PID控制器参数变化曲线的仿真结果。基于实验室挖掘机实验平台,编写各控制算法的控制程序,进行相关的实验研究,将普通PID控制器,变速积分PID控制器、单神经元自适应PID控制器以及BP神经网络PID控制器的实验结果进行分析和比较。实验结果表明,基于神经网络的单神经元自适应PID控制器和BP神经网络PID控制器控制效果比常规PID控制器控制效果要好,适应性更强,而两者中BP神经网络PID控制效果更佳,具有很好的应用前景。
[Abstract]:The excavator has been widely used in the field of construction because of its high performance and high efficiency. However, most of the excavators depend on manual operation of the excavator in the current operation process, and it is not suitable to work in dangerous environment or for a long time. It is very important to realize the automation and intelligence of excavator. Taking the experimental platform of laboratory excavator as the research object, the purpose of this paper is to design a kind of high performance controller to realize the high performance control of the track of the excavator's working device. This paper mainly includes the modeling of excavator tool trajectory control system, the design and simulation of controller and the experimental research. In the process of modeling the control system of excavator, the control system of moving arm and bucket rod is modeled separately, and the transfer function of each link is calculated. The dynamic mechanism of hydraulic system models the forward and backward movement of piston in hydraulic cylinder, and the transfer function is obtained by least square fitting in mechanical structure. In order to realize digital computer control, the continuous transfer function is discretized and its pulse transfer function and difference equation are obtained. Based on the discrete mathematical model and incremental PID control algorithm of excavator tooling trajectory control system, The variable speed integral digital PID controller based on ordinary PID controller and the single neuron adaptive PID controller and BP neural network PID controller are designed. The control algorithm program of the controller is written in the MATLAB software, the parameters of the controller are set and simulated, and the control system of the moving arm and the bucket rod control system are obtained under the action of each controller. The simulation results of the response curve of the test signal and the parameter change curve of the PID controller are given. Based on the experimental platform of the laboratory excavator, the control program of each control algorithm is compiled, and the related experimental research is carried out. The general PID controller and the variable speed integral PID controller are used. The experimental results of single neuron adaptive PID controller and BP neural network PID controller are analyzed and compared. The experimental results show that the control effect of single neuron adaptive PID controller and BP neural network PID controller based on neural network is better than that of conventional PID controller, and the BP neural network PID control effect is better than that of conventional PID controller. It has a good application prospect.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TU621;TP183;TP273
，
本文编号：2311842
[Abstract]:The excavator has been widely used in the field of construction because of its high performance and high efficiency. However, most of the excavators depend on manual operation of the excavator in the current operation process, and it is not suitable to work in dangerous environment or for a long time. It is very important to realize the automation and intelligence of excavator. Taking the experimental platform of laboratory excavator as the research object, the purpose of this paper is to design a kind of high performance controller to realize the high performance control of the track of the excavator's working device. This paper mainly includes the modeling of excavator tool trajectory control system, the design and simulation of controller and the experimental research. In the process of modeling the control system of excavator, the control system of moving arm and bucket rod is modeled separately, and the transfer function of each link is calculated. The dynamic mechanism of hydraulic system models the forward and backward movement of piston in hydraulic cylinder, and the transfer function is obtained by least square fitting in mechanical structure. In order to realize digital computer control, the continuous transfer function is discretized and its pulse transfer function and difference equation are obtained. Based on the discrete mathematical model and incremental PID control algorithm of excavator tooling trajectory control system, The variable speed integral digital PID controller based on ordinary PID controller and the single neuron adaptive PID controller and BP neural network PID controller are designed. The control algorithm program of the controller is written in the MATLAB software, the parameters of the controller are set and simulated, and the control system of the moving arm and the bucket rod control system are obtained under the action of each controller. The simulation results of the response curve of the test signal and the parameter change curve of the PID controller are given. Based on the experimental platform of the laboratory excavator, the control program of each control algorithm is compiled, and the related experimental research is carried out. The general PID controller and the variable speed integral PID controller are used. The experimental results of single neuron adaptive PID controller and BP neural network PID controller are analyzed and compared. The experimental results show that the control effect of single neuron adaptive PID controller and BP neural network PID controller based on neural network is better than that of conventional PID controller, and the BP neural network PID control effect is better than that of conventional PID controller. It has a good application prospect.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TU621;TP183;TP273
，
本文编号：2311842
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2311842.html
