当前位置:首页 > 新闻报道

实验室杨世锋硕士研究生毕业答辩

发布时间:2016-06-03    字体:[增加 减小]  

答辩人:杨世锋 硕士研究生

学科专业:物流工程

论文题目:基于WSN的多移动机器人控制系统研究

摘要:

在信息时代背景下,随着物联网技术的兴起,无线传感网络和机器人技术在不同的领域得到越来越广泛的应用。无线传感网络具有低功耗、自组织等特性,机器人技术具有智能化、适应性等特点,两者结合运用产生了集传感器网络、自动控制、通信技术、网络反馈等技术于一体的无线传感网络控制平台,具有很高的应用研究价值。在军事仿真、智能交通、工业生产、物流运输等领域有广泛的应用前景。

针对多机器人执行任务的轨迹跟踪问题,本文通过理论研究、框架设计、系统集成等过程构建了一个基于无线传感网络反馈的多移动机器人控制系统。该控制系统在方法上融合了无线传感网络和移动机器人技术,从而解决了工程项目中连续定位数据反馈下的移动机器人轨迹跟踪问题,实现了对多移动机器人的移动控制。

首先,本文从系统总体方案设计出发,对功能需求进行了分析,描述了构建无线传感网络控制系统的关键技术,搭建了系统的初步框架。控制平台由视觉定位、控制计算机、移动机器人三部分组成。研究从三个部分的硬件选型和软件工作流程等工作着手,对系统进行设计分析。为具体集成和实现打下基础。

其次,由于无线传感网络是系统信息连接交互的基础,为了保证指令传达的有效、及时,本系统采用了ZigBee无线传感网络作为通讯方式,并介绍了其组网方式、数据收发指令协议和传输模式等。对移动机器人在笛卡尔坐标系下运动学模型进行了分析,在移动机器人控制端采用模糊控制算法,以偏角控制为例设计了模糊控制器,对移动机器人进行嵌入式编程实现基于策略调整的任务指令执行流程,完成了实验验证前的准备工作。

最后,在系统平台集成和控制算法实现的基础上,在5000mm*2400mm的场地中设计了多移动机器人控制系统的实验,进行了实验数据的采样收集。为了评价系统的跟踪精度,设计了基于最小二乘法的拟合函数偏差计算方法,运用Matlab对实际轨迹与跟踪轨迹进行分段拟合计算,最终得出偏离理想轨迹的平均误差,结果分析证明本机器人控制系统能够有效的跟踪理想轨迹,为多机器人控制系统进一步的研究和应用提供了实践支持。

Abstract:

With the development of Internet of Things technology, Wireless Sensor Network and robotics are applied more and more extensively in different fields in the information era. Wireless sensor network has the characteristics of low power consumption, self-organization, and robotics has the traits of intelligence, adaptability. The appliance of both two technologies generates a control platform based on wireless sensor network, which combined sensor network, automatic control, communication technology and network feedback control, creates great practical and researching value. It is expected to own a wide application prospect in military simulation, intelligent transportation, industrial production, logistics, transportation and other fields.

With the method of integrating Wireless Sensor Network and mobile robotics, in accordance with trajectory tracking problems in the executions of mobile robots, in this article, a multi-robot control system based on feedback of wireless sensor network will be presented, which will solve mobile robot trajectory tracking control problem under continuous positioning data feedback in engineering project, realized the mobile control of multiple mobile robots.

Firstly, this article embarks from overall scheme design of the system, analyzes the functional requirements, sets up a preliminary framework for the system, and describes the construction and the key technologies of wireless sensor network control system. Control platform consists of three parts: vision localization, center computer and mobile robots. The analysis and design the whole system will be commenced from hardware selection and software process, which will establish a foundation for the specific integration and implementation.

Secondly, as the wireless sensor network is the basis of system information interaction, in order to ensure the transferring of instructions being effective and timely, this system employs the ZigBee technology as a way of communication, introduces the network mode, data transceiver command protocol and transmission mode, etc. The article will analyze the kinematics model of mobile robots in cartesian coordinate system, using fuzzy control as the mobile robot control algorithm. An angle fuzzy controller will be designed as an example, embedded programming implements the task instruction execution process which based on strategy adjustment, the preparation work is supposed to be completed before the experiment.

Finally, on the basis of integration of system platform and control algorithm, an experiment of multiple robot control system in a field of 5000mm*2400mm will be designed, for acquiring data and collecting samples. To evaluate tracking precision of the system, a fitting function deviation calculation method based on the least squares will be employed, the Matlab will be applied to process piecewise fitting calculation between actual trajectory and tracking trajectory, the average deviation from the ideal trajectory will be observed eventually, the average error analysis results will prove that the robot control system can track the ideal trajectory effectively, which would provide some practical supports for further research and application of multiple robot control system.