基于ARM-Linux的爬壁机器人远程自动控制系统  

Remote automatic controller of climbing robots based on ARM and Linux

作　　者：何宏[1] 赵磊[1] 张志宏[1,2] He Hong Zhao Le Zhang Zhihongla Tianjin(Key Laboratory of Complex System Control Theory and Application, School of Electrical, University of Technology, Tianjin 300384, Chin Transmission and Launch Department, Tianjin Radio and TV Station, Tianjin 300072, China)

机构地区：[1]天津理工大学自动化学院天津市复杂系统控制理论及应用重点实验室,天津300384 [2]天津广播电视台传输发射部,天津300072

出　　处：《国外电子测量技术》2016年第12期56-60,共5页Foreign Electronic Measurement Technology

基　　金：天津市科技支撑重大科技工程专项基金项目(14ZCDGSF00028);"天津市高等学校创新团队培养计划资助"项目(TD12-5015)资助

摘　　要：目前,船舶清洗系统的爬壁机器人普遍采用无线工控机进行控制,为了改进无线工控机不能实时显示现场工况和状态信息以及自动寻径控制的缺点,研制了一种控制系统以实现控制的远程可视化,并能自动寻径。系统使用ARM-Linux平台,外部电路包括摇杆电路、串口-ZigBee电路、视频解码电路等。自动控制算法采用基于图像识别的跟踪迹Q(λ)算法,能够通过图像实现算法快速收敛,完成自动寻径。经过测试,系统能实现对爬壁机器人的自动控制,并且跟踪迹Q(λ)算法比传统的Q-learnig算法或Q(λ)算法收敛速度分别提高22.8%和13.8%。At present,the robot of the ship cleaning system is controlled by the wireless industrial controller in common.To solve the problem of that the wireless controller cannot display the working condition or route automatically,a control system which can display the working station remote and route automatically.This system,based on ARM and Linux,includes rocker circuit,serial-port ZigBee circuit and video recording circuit.The automatic control algorithm,the tracking Q（λ）algorithm based on image recognition,is to control the robot tracking automatically.From the test results,this system can control the robot automatically,and the speed of convergence of tracking Q（λ）algorithm is 13.8% faster than Q-learning algorithm or 22.8%than Q（λ）algorithm.

关 键 词：爬壁机器人 LINUX 远程控制 可视化 

分 类 号：TN108.4[电子电信—物理电子学]