To effectively manage the flow of unmanned vehicles in large manufacturing plants and warehouses, both in open spaces and indoors, an accurate dynamic positioning system is required. This is necessary to optimize routes in order to increase the speed of movement and prevent collisions. The article considers the application of an autonomous navigation system that uses radar motion sensors together with odometric data. The developed algorithm of adaptive data filtering based on the Kalman filter is described, which allows continuously receiving information about the current position of unmanned vehicles with the required accuracy. As a result, it is possible to optimize the trajectory of all unmanned vehicles in dynamics in order to increase the productivity and reliability of cargo transportation.