Towards a Conceptual Design of Intelligent Material Transport Using Artificial Intelligence

Reliable and efficient material transport is one of the basic requirements that affect productivity in industry. For that reason, in this paper two approaches are proposed for the task of intelligent material transport by using a mobile robot. The first approach is based on applying genetic algorithms for optimizing process plans. Optimized process plans are passed to the genetic algorithm for scheduling which generate an optimal job sequence by using minimal makespan as criteria. The second approach uses graph theory for generating paths and neural networks for learning generated paths. The Matlab software package is used for developing genetic algorithms, manufacturing process simulation, implementing search algorithms and neural network training. The obtained paths are tested by means of the Khepera II mobile robot system within a static laboratory model of manufacturing environment. The experiment results clearly show that an intelligent mobile robot can follow paths generated by using genetic algorithms as well as learn and predict optimal material transport flows thanks to using neural networks. The achieved positioning error of the mobile robot indicates that the conceptual design approach based on the axiomatic design theory can be used for designing the material transport and handling tasks in intelligent manufacturing systems.