Development of proposals for formalising the process of filtering navigation information of an underwater robot at shallow depths

Abstract

The purpose of the article is to increase the efficiency of solving the problems of stabilising underwater robots at shallow depths based on complex processing of navigation information and its filtering using the Kalman theory. This goal is achieved by defining a set of mathematical dependencies for formalising the process of filtering navigation information of underwater robots based on complex data processing. In this case, the filtering is carried out using a distributed set of Kalman filters of different structures, which were selected considering the characteristics of the data being evaluated. It has been established that at present, underwater robots at shallow depths are widely used around the world for various tasks, including search operations and underwater inspections. However, the operation of these robots is characterised by difficult conditions. These conditions include unknown parameters of underwater navigation, the impact of external disturbances, changes in the mass, size and hydrodynamic characteristics of robots while operating in water. Currently, the concept of control based on intelligent methods is considered a promising approach to automating the control of moving objects. However, the use of such controllers for underwater robots, together with the problems of obtaining up-to-date navigation information, has not yet achieved sufficient efficiency. In addition, the issues related to the development of a navigation information processing system using nonlinear filters and the creation of intelligent controllers for underwater robots are still insufficiently covered in the scientific and technical literature. The most significant result is a set of mathematical dependencies for formalising the process of filtering navigation information of underwater robots using a set of distributed Kalman filters of different structures. Such sets are closely correlated with the relevant characteristics of the data being evaluated. In this context, the inertial module with Kalman filtering algorithms can be used to measure angular motion parameters and solve the problems of roll, pitch and yaw stabilisation. Due to the low speeds of underwater robots at shallow depths and the absence of high-frequency interference in the pressure sensor measurements, the data from the pressure sensor can be used to determine the vertical speed