Design And Implementation Of Self-tracking Robot

At present, manual operation is used in the most adverse circumstancesanddangerous environment of the industrial scene, however, its high-risk, poorsecurity lead to the urgent need for intelligent guided robot researchanddevelopment. And now, the research development on guiding robot ismostlybased on the MCU and DSP, so there is a lot of limitation, such as therelativesingle function, weak expansibility, and so on. Thus it can not maketheguided robot accomplish more and more complicated tasks withhigherintellectualized character.Therefore, this paper based on the intelligentcharacter of robot, adoptsmore advanced embedded MCU as ahost-controlled chip, makes transplantedLinux operating system as a robot’sbrain, and research the speediness andstability of the traveling automaticline-tracking robot on the black tracking lineof the white course. It has constructed the intelligent robot’s visual guidancesystem,designing and developing each hardware module of the robot.Modelanalysis to turning of servo, based on the robot’s body axisrelative to thedirection and lateral deviation of the path midline as inputdesign a fuzzycontroller for the turning of servo. Model analysis to robot’sspeed controldesigns a classical PID controller and a single neuralnetwork-based PIDcontroller. It has successfully transplanted Linux2.6kernel to S3C2440corechip,meanwhile, transplanted a matched BIOS startup program UBOOT andmadethe root file system JFFS2, and has accomplished drive andapplicationprocedures based on Linux2.6kernel according to the robot’shardware design,and in the end, made testing, comparing and analysis to the selfline-trackingrobot by various calculating methods.This paper has used embedded Linux system to the intelligent guidedrobot.It is proved by experiments that the robot can accomplish automaticline-tracing task fleetly and stably after input the road surface informationcollected by theCCD senor. It is also proved that robot’s speed can becontrolled well by singleneural network PID controller in curve paths, according to the testing andcomparing to the two calculation methods ofrobot’s speed controlling on realitycondition.