基于智能小车的模拟入库系统

发布时间：2018-03-06 18:27

本文选题：Arduino　切入点：Enhanced　出处：《西安科技大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着工业自动化的高速发展，传统的人工入库已不能满足仓储自动化发展需求。本文的设计是以深圳市中科鸥鹏智能科技有限公司提供的基于Arduino Enhanced Board控制板的车载机械臂智能小车为实验平台，针对传统的人工入库存在的问题，设计了基于智能小车的模拟入库系统，基于智能小车的模拟入库系统实现的功能是小车巡线行走，到达取货点后超声波远距离测量物体位置，由机械臂抓取物体，然后进行颜色识别并放置到与物体颜色对应的位置完成入库，最后小车回到初始位置。论文是基于智能小车的模拟入库系统的软件设计，同时配合硬件的电路，完成了整体系统的制作，并对软、硬件进行了联机调试，为深圳市中科鸥鹏智能科技有限公司开发了第一代可行走的机械臂小车。硬件模块包括Arduino Enhanced Board控制板、巡线模块、超声波模块、颜色识别模块、机械臂模块；软件模块包括巡线模块、超声波测量模块、颜色识别模块以及机械臂控制模块。硬件方面首先对各个模块所用到的器件进行选型、分析与比较，针对不同的器件设计相应的驱动电路，然后对各个硬件模块进行设计，并对各个硬件模块的功能进行测试，最后将所有硬件模块整合在一起进行调试。软件方面主要是针对各个硬件模块编写对应的代码，其中重点是颜色识别模块和机械臂控制模块软件设计。颜色识别模块是针对传统的阈值法识别精度较低、鲁棒性较弱的问题以及基于神经网络的最小色差法计算量大、难以应用的问题，提出了基于TCS230传感器的模糊识别颜色算法。算法实现过程是首先根据少量的标准样本确定出模糊关系，将TCS230传感器实时测得的RGB数据模糊化，根据离线求出的模糊关系解出模糊输出，通过解模糊确定出识别的物体颜色。模糊识别算法不仅识别精度较高，计算量较少，而且鲁棒性较强，能够应用到很多自动化系统。机械臂模块是采用传统的D-H法离线对机械臂进行建模，求出变换矩阵，通过对矩阵齐次变换得到逆解方程，求解方程后将结果写入程序以控制机械臂。D-H法相对于遗传算法等智能算法计算量较少，而且易于实现，适合应用在本文设计的智能小车系统中。智能小车模拟入库的100次实验结果证明，设计的智能小车模拟入库系统对物体颜色识别率高，搬运精确。同时系统的安全性较高，，而且响应也较快，为仓储的入库系统的发展提供了参考。
[Abstract]:With the rapid development of industrial automation, The traditional manual storage can no longer meet the requirement of warehouse automation development. The design of this paper is based on the Arduino Enhanced Board control board, which is provided by Shenzhen Zhongke Gupeng Intelligent Technology Co., Ltd. In view of the problems existing in the traditional manual storage system, an analog storage system based on intelligent trolley is designed. The function of the simulation storage system based on intelligent trolley is to inspect the track and measure the position of the object at a long distance by ultrasonic wave. The object is grabbed by the robot arm, then color recognition is carried out and placed in a position corresponding to the color of the object. Finally, the vehicle returns to the initial position. The paper is based on the intelligent car simulation system software design, and with the hardware circuit, completed the production of the whole system, and the software and hardware on-line debugging. For Shenzhen Zhongke Gupeng Intelligent Science and Technology Co., Ltd, the first generation walking robot is developed. The hardware module includes Arduino Enhanced Board control board, line patrol module, ultrasonic module, color identification module, robot arm module; The software module includes inspection module, ultrasonic measurement module, color identification module and manipulator control module. In the hardware aspect, the devices used in each module are selected, analyzed and compared, the corresponding driving circuit is designed for different devices, and then each hardware module is designed, and the function of each hardware module is tested. Finally, all the hardware modules are integrated together for debugging. In the software aspect, the corresponding codes are mainly written for each hardware module, in which the emphasis is on the software design of the color recognition module and the robot arm control module, and the color recognition module is aimed at the low accuracy of the traditional threshold method. The problem of weak robustness and the problem that the minimum chromatic difference method based on neural network is difficult to be applied, A fuzzy color recognition algorithm based on TCS230 sensor is proposed. Firstly, the fuzzy relation is determined according to a small number of standard samples, and the RGB data measured in real time by TCS230 sensor is fuzzied. The fuzzy output is solved according to the fuzzy relation obtained from the off-line, and the object color is determined by the fuzzy solution. The fuzzy recognition algorithm not only has higher recognition accuracy, less computation, but also has stronger robustness. It can be applied to many automation systems. The traditional D-H method is used to model the manipulator off-line, and the transformation matrix is obtained, and the inverse solution equation is obtained by homogeneous transformation of the matrix. After solving the equation, the results are written into the program to control the manipulator. D-H method has less computation than the intelligent algorithm such as genetic algorithm, and it is easy to realize. It is suitable for application in the intelligent car system designed in this paper. The results of 100 experiments show that the designed intelligent vehicle simulation storage system has high recognition rate of object color and accurate handling. At the same time, the security of the system is high and the response is fast. It provides a reference for the development of warehousing system.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TH692.3

【参考文献】