The study focuses on the development of an integrated navigation system for mobile platforms operating indoors and in urban environments where satellite signals are unavailable or intentionally distorted. The aim of the research is to develop an algorithm for deep integration of inertial measurements, visual odometry, and radio beacon data with simultaneous estimation of observability and sensor drift parameters, as well as experimental validation of the proposed solution. The materials include results of numerical simulations of the trajectories of a moving object and experimental data obtained from a robotic platform equipped with an inertial module, a stereo camera, and a system of stationary ultra-wideband radio beacons at a closed test site with a complex route configuration. The methodology is based on an extended Kalman filter with a fifteen-dimensional state vector that includes coordinates, velocity, orientation, and quasi-stationary errors of accelerometers and gyroscopes. A formal observability analysis is performed by examining the rank of the matrix, and the influence of trajectories and beacon geometry on drift identifiability is investigated. The results of simulations and experiments show that a purely inertial system leads to an avalanche-like growth of positioning error, whereas the addition of visual odometry and radio beacons exponentially reduces the root mean square error of absolute positioning down to tenths of a meter along the entire route. It is shown that the integrated configuration provides high accuracy in estimating inertial sensor biases with a relative error on the order of a few percent, which significantly slows the accumulation of error during temporary loss of external measurements and ensures smooth degradation of accuracy in case of subsystem failures. The study additionally demonstrates a critical dependence of accuracy on the geometric configuration of radio beacons and formulates engineering recommendations for their placement to achieve the target tolerance level. The scientific novelty of the work lies in the simultaneous quantitative consideration of observability of drift parameters and the influence of infrastructure geometry, which makes it possible to define justified requirements for platform maneuvers and sensor configuration already at the system design stage. Future research directions include the use of more general nonlinear filters, analysis of system performance under non-Gaussian noise conditions, and the development of adaptive procedures for adjusting measurement weights based on the current informational value of signals.