Levkivskyi R. Models and methods of dynamic joint motion planning for unmanned vehicles during industrial fishing

The purpose of the study is to improve the quality, efficiency, and safety of industrial fishing operations performed by a group of unmanned vehicles based on models and methods of dynamic planning of safe trajectories of their joint motion. Research methods: methods of set theory, graph theory and control theory to build a model of joint operation; methods of topology theory to build a model of interaction space, the formal apparatus of theories of fuzzy, rough and soft sets to represent the uncertainty of information; methods of fuzzy, rough and soft sets and their combinations to build a configuration space; methods of similarity theory for building a hybrid trajectory search model; methods of situational control theory to build a model for finding trajectories by cases; methods of constraint satisfaction for the synthesis of spatial solutions using a hybrid model of trajectory search; machine learning methods for organizing a common spatial model; adaptive planning methods to solve the problem of trajectories planning. First developed a hybrid method for finding safe trajectories in a configuration space represented by a soft rough or soft fuzzy-rough topology using a safe motion corridor based on the superposition of multilevel collision cone systems and superimposed on a soft interaction space topology that consistently narrows the configuration space using heuristics based on the nonlinear properties of the search space. The model of knowledge representation about dynamics of interaction of a group of unmanned vehicles during their joint movement has been further developed based on the use of multilevel safety domains, represented by a soft rough spherical topology in the form of concentric spatial regions, which allows to determine non-spherical safety domains of any size and shape by counting different radii at different longitude and latitude. The information technology to solve the problem of dynamic planning of safe trajectories of a group of unmanned vehicles has been improved, which is based on a hybrid method of finding safe trajectories in a configuration space represented by a soft rough topology that allows to combine methods of artificial potential fields, definition of similar spatial configurations and RRT. Field of use - transport and special systems, which use group control of the joint motion of moving objects of any type.