Neural algorithms of path planning for mobile robots in transport systems

Abstract

Methods of control of mobile robotic systems are described. To solve the problem of real time motion planning and control of several mobile robots, it is necessary to apply methods of neural control, combined with artificial intelligence (AI). The control system may be of single-level type (with equal robots) and two-level type with a special coordinator (supervisor), which provides motion coordination for individual robots. The main and most difficult problem of optimal control of a robot team is to plan robots motion, concerning both obstacles in the working zone and other moving robots. In mathematical sense the problem can be considered as optimization of the whole system functional which consists of individual robot functionals with constraint of some common resources. For a group of mobile robots such resources are represented by a free space in the working zone.