Solving the Problem of Autonomous Navigation via the Integration of Inertial and Optical Navigation Systems

Currently, when solving the problem of autonomous navigation by the parameters of an optical flow recorded by a video camera during the motion of an object, only the components of the linear and angular velocities of an object are measured. Such estimation of velocities is only a pert of the general navigation problem–current object positioning and angular orientation determination–and provide no solution for this problem as a whole. In this connection, the integral approach combining the capabilities of an inertial navigation system, which makes it possible to solve the autonomous navigation problem as a whole, and a system of navigation by an optical flow, which provides the autonomous observation of linear and angular motion parameters at minimum hardware expenditures, is considered. Since, a serious problem encountered in the use of these systems is the difficulty of taking into account noises of different probabilistic nature, the synthesis of a stochastic model for the proposed integrated inertial-optical system is carried out with allowance for the further application of methods, which take into consideration the effect of noises in the estimation of object navigation parameters, i.e., the methods of contemporary stochastic filtering theory. As a result, to estimate the full vector of object motion parameters by integrated inertial-optical navigation system measurements, a modified extended Kalman filter taking into account the correlation between object and observer noises has been constructed. A numerical experiment illustrating the efficiency of the proposed approach has been performed.