The vector control system of an asynchronous motor, which is closed in terms of the rotor speed, assumes the presence of a sensor on the shaft of the electric machine. However, in practice, often there are problems of regulating the speed of rotation of asynchronous electric drives, in the solution of which the use of mechanical movement coordinate sensors is technically irrational or impossible. In this case, it is necessary to use one of the sensorless vector control algorithms, which involve direct measurement of only electrical quantities. In rotor vector control systems, when the stator is connected to the network at the same time, the need to measure the angle of rotation of the rotor relative to the stator is added, which complicates the structure of the observer.
 For the synthesis of speed observers in asynchronous electric drives, the approach known in control theory is widely used, in which, based on the second Lyapunov method, an adaptation function is formed that provides asymptotic convergence of the adaptive model to the reference one.
 In this paper, the synthesis of the rotor speed observer of an asynchronous gate cascade (AGC) was carried out for its application in a sensorless relay-vector control system. Equations of electromagnetic processes in an asynchronous machine, formulas for coordinate transformations and an identifier for the position angle of the rotor AGC were used as initial mathematical models. The observer synthesis is based on the second Lyapunov method, which allows to determine the structure of the adaptation function, which adjusts the adaptive model in such a way that the difference between the outputs of the adaptive and reference models tends to zero. Thus, the adaptation function, which uses the stator flux vector, ensures the asymptotic stability of the perturbed motion. The proposed speed observer AGC differs from the known observers for stator control in that, along with the speed calculation, the trigonometric functions of the rotor rotation angle are determined, and the structure of the observer includes coordinate transducers that are absent in stator vector control systems, since they do not need to calculate the projections of the voltage vector of one part of the machine onto the orthogonal axes of its other part.