Abstract:
An AGV (Automated Guided Vehicle) is one example of robotic systems that can move freely on a guidance of predetermined tracks. Meanwhile AMR (Autonomous Mobile Robot) is another system that is capable of moving autonomously without following a
predetermined tracks or any human intervention and is also capable of adapting itself to a dynamical change of its environment. Hence, this makes it differ from and more difficult to be controlled than the AGV. Level of AMR adaptabilty much depends on the completeness of sensor system planted in it. This paper will describes what is called world modeling,
where robot behaviour performs a modeling task in which the robot resides. This task is very useful for instance in determining a completely unknown layout of a new plant, hazardous or dangarous places for human being and so on. Moreover, this behaviour is suitable for mapping task in the IEEE micro mouse competition.
This research focuses on the design of an AMR prototype and the development of motion algorithm for mapping an unknown maze. A two stepper motor AMR has been developed equipped with four IR sensors to detect any wall in its surounding. The maze consists of
several 18cm x 18cm cells that is separated by walls according to the IEEE Micromouse Rules. The robot needs a navigation system which consists of sensor, motion and control systems. An algorithm based on depth first search (DFS) and A star concepts has been
developed to control the motion of the robot when mapping the unexplored maze. Result shows that mapping the unkown maze has been successfully done.
Combination of DFS + A star algorithm proves itself to be more effective when compared to pure DFS concept.
