Intelligent Control for Container Terminal AGV

Abstract

The world's container cargo trading is increasing daily, and the role of the container terminal is becoming more important as the center of cargo transportation.
In Japan, new container terminals being constructed face very severe competition with larger, cheaper terminals so they must handle cargo more efficiently and cheaply. To cope, handling systems such as container cranes, yard cranes, and conveyers are being automated to enable unattended operation unloading and loading schedule planning, etc., are being computerized.
In these system, crane automation and control computerization are already generalized, but automatic control of container conveyers is not completed yet.
The container conveyer -15m long, 4m wide, and 1.5m high - picks up containers from container ships with a container crane and hauls them to the shift yard for release to the yard crane. Both crane are operated automatically, so the conveyer must stop at the desired position within a permissible error of ±7.5cm, and run on a predetermined course and speed. Collision avoidance is required because many vehicles oparate on the same course.
The automated guided vehicle (AGV) system, which is diesel-driven, 4WS and 4WD, was thought to be effective, but container weights very widely, as do road conditions which depend on weather, so conventional control is not sufficient to maintain the required accuracy.
We applied learning control to maintain the desired course and for stopping at the desired position. Speed was controlled, conventionally.
The system's applicability was confirmed with computer simulation and vehicle performance testing.
This system will be used at the Kawasaki container terminal in 1999 and we are currently working on improving performance.