DEVELOPMENT OF A HYBRID METHOD FOR PLANNING THE MOVEMENT OF A GROUP OF MARINE ROBOTIC COMPLEXES IN AN A PRIORI UNKNOWN ENVIRONMENT WITH OBSTACLES

Abstract

The article discusses the problem of organizing the group movement of marine robotics
(MRS), in particular, unmanned autonomous underwater vehicles (AUV), in an a priori unknown
environment with obstacles. A brief analysis of existing projects on the subject of MRS group control,
and path-planning algorithms have been carried out. The presence of numerous studies in
this area confirms the relevance of the indicated problem. The formal statement of the problem of
the movement of four robots in formation is presented. The proposed method for a group path
planning is based on a combined approach that organizes a multi-level solution to organizing the
movement of MRS. At the upper level, a system for global path planning and mission control was
developed based on the random tree method, which provides the general movement of the group
based on a priori information about the state of the environment. The lower-level planning system
adjusts the global path, allowing objects at the local level to carry out the group agents’ movement
and interaction, including ensuring their collision-free movement in environment and exit from the
areas of local minima. The article provides a detailed analytical description of the developed algorithm
and a block diagram of its functioning. Numerical simulation of the movement of a group
of 4 AUVs in a non-deterministic environment with fixed obstacles is carried out. The simulation
was carried out taking into account obstacles of various shapes and complexity. The results of
mathematical modeling demonstrated the solution to the problem of the exit of the AUV group
from the area of the local minimum. Full-scale tests are presented on the example of a group of
three unmanned boats: a group of mobile objects formed a formation and carried out a movement
in a given formation to target positions and returned to the final zone. In addition, the local planning
module developed within the framework of this article was integrated into the software of the
"Shadow" underwater glider control system. In the conclusion, the results of the proposed method
and its further development, in particular, its application in 3D environment, are considered.