A Navigation Control Algorithm for Automated Guided Vehicle Based on Neural Network Sensing Prediction

신경망 예측에 기반한 AGV의 주행 알고리듬

  • 나용균 (포항공대 전자전기공학과) ;
  • 김선효 (포항공대 전자전기공학과) ;
  • 오세영 (포항공대 전자전기공학과) ;
  • 성학경 (삼성전자 생산기술연구소) ;
  • 김성권 (삼성전자 생산기술연구소)
  • Published : 2000.10.01

Abstract

A robust intelligent algorithm for AGV navigation control is presented here based on both magnetic and gyro sensors to track a reference trajectory. Since the proposed system uses an intermittent array of short magnetic tape strips, it lends itself to a very easy installation and maintenance compared to other types of positioning references such as electric wire, magnets, RF and laser beacons. The neural network is to predict the lateral deviation of the AGV in the intervals where no magnetic tape references are available. Further, the use of intelligent control ensures a robust and flexible control performance. Computer simulation of AGV control demonstrates its adequate tracking performances even where the sensor information is not available. Real experiments using Samsung AGV are also on the way for real verification

Keywords