Home – Home – Jornals – Avtometriya 2025 number 6

Jamming, as well as various types of spoofing of signals from satellite navigation systems, which are currently the most effective tools in terms of positioning accuracy for moving objects, create a need to search for alternative navigation methods that ensure both accuracy and autonomy in solving the navigation problem. Among such approaches, various methods of integrating inertial navigation systems (INS) with optical navigation systems (ONS) can be highlighted, as they ensure the maximum accuracy over long intervals of motion and the minimum cost. For objects moving along arbitrary trajectories (in the absence of terrain maps and reference points), the existing methods of processing information from optical navigation systems, obtained in the form of an optical flow velocity field, enable the estimation of the current object's velocity projections - both linear and angular. However, the application of traditional methods, firstly, does not allow solving the navigation problem as a whole (determining the object's coordinates and its spatial orientation angles) and, secondly, leads to computational costs that are often critical for onboard computers. In this regard, the article considers an approach that combines the capabilities of an inertial navigation system, which provides a comprehensive solution to the autonomous navigation problem, and a developed optical observer of the object's motion parameters. This observer allows for autonomous estimation of navigation parameters without preliminary calculation of the velocity field, i.e., with the minimal computational costs. For a more complete coverage of the problem, cases of rigid camera mounting on the object, as well as its two- and three-axis stabilization, are considered. The final synthesis of algorithms for stochastic estimation of the object's motion parameters, which accounts for the uncertainty of the probabilistic characteristics of noise in a real inertial-optical navigation system, is carried out considering the possibility of subsequent application of methods from the modern nonlinear filtering theory. Due to specific features that arose during the development of the considered inertial-optical navigation system - namely, the correlation between the object and observer noises - a variant of the extended Kalman filter for correlated noises is used for stochastic estimation of the navigation parameter vector. The results of numerical modeling of the proposed navigation algorithm illustrate the possibility of its effective practical application.