Development of a Cascade Intelligent System for Path Planning of the Group of Marine Robotic Complexes

Abstract

Artificial Intelligence (hereinafter referred to as AI) systems have recently found great application and use in various industries, such as data processing, data analysis, and the operation control of marine robotic complexes, etc. In view of the ever-increasing degree of complexity of the missions assigned to marine robotic systems, it is becoming obvious that the AI technologies should be used as combined systems which can provide control of marine robotic complexes (hereinafter referred to as MRCs), their navigation in sea, logic formation of MRC behaviour in uncertain environments, path planning, and processing optimization of the received MRC payload data. All the areas mentioned above are within the field of MRC development, and currently do not have a general solution. This article discusses the development process of an intelligent system for path planning of a group of marine robotic complexes. The architecture of the intelligent system is based on a cascade approach, which includes the consistent use of functional modules designed as various “organs of perception” of the system. A detailed description of the development of each module and mathematical modelling of the presented algorithms are provided in this paper, and the main results of the conducted full-scale experiments are demonstrated.

Keywords

• neural network
• machine learning
• reinforcement learning
• DQN
• group control system
• path planning