To ensure the stabilization of the equilibrium state in a nonlinear system in the presence of noise, it is not enough to solve the local stabilization problem, it is also necessary to ensure continuous monitoring of a possible transition to a critical state leading to system failure. To organize such monitoring, we use the large deviations principle applied to dynamical systems with small perturbations. For the purposes of monitoring, the optimal path that we call the A-profile is important. We use the A-profile to build a situational forecast in the risk control problem for a multi-agent system. In addition to the nonlinear mechanism of internal stabilization of the level h for each of the agents, there are forces of mean field interaction between the agents. The weak limit in this model with the number of agents tending to infinity is described by the FokerPlanck-Kolmogorov equation, but the use of approximation up to O(h²) leads to a finite-dimensional Wentzel-Freidlin scheme. According to the scheme, we obtain an explicit A-profile as a solution of the degenerate Abel equation of the second kind. At the same time, the approximation in h makes it possible to develop a method of successive approximations for the A-profile. In this paper, the A-profile is synthesized as a solution of the optimal control problem, where the state-dependent Riccati equation method and the method of the approximating sequence of Riccati equations are used. In the article, these methods are applied and compared within the framework of the risk control problem.

The issues of expediency of using adaptation tools in robust control systems are discussed. It is stated that purely robust systems without the use of adaptation tools, in principle, cannot provide high efficiency and versatility in solving practical problems. Robust control systems were originally conceived as non-adaptive systems (passive adaptive systems). In view of the emerging problems in solving practical problems, then a mutation occurred and works appeared under the name "Adaptive-robust systems", "Combined robust systems", etc., partially using adaptation algorithms. Setting the problem of control synthesis under conditions of uncertainty without elements of adaptation is figuratively speaking similar to "search for a black cat in a dark room, especially if it is not there" (Confucius). The most adequate from the point of view of compliance with the fundamental principles of the theory of automatic control is an approach based on an increase in the gain of an open loop. Nevertheless, here too a problem arises — an increase in the gain violates the stability of a closed system. All known research is concentrated around the solution of this problem. An integrated gain self-tuning algorithm and the corresponding circuitry and Simulink implementation scheme have been developed. The reliability of theoretical reasoning was verified by simulating a limiting robust system with self-tuning and a parametrically indeterminate object — "peak gyroscope (parametric pendulum)". Computer studies have made it possible to draw a number of positive conclusions that are of great practical importance. 

Reinforcement learning is based on a principle of an agent interacting with an environment in order to maximize the amount of reward. Reinforcement learning shows amazing results in solving various control problems. However, the attempts to train a multitasking agent suffer from the problem of so-called "catastrophic forgetting": the knowledge gained by the agent about one task is erased during developing the correct strategy to solve another task. One of the methods to fight catastrophic forgetting during multitask learning assumes storing previously encountered states in, the so-called, experience replay buffer. We developed the method allowing a student agent to exchange an experience with teacher agents using an experience replay buffer. The procedure of experience exchange allowed the student to behave effectively in several environments simultaneously. The experience exchange was based on knowledge distillation that allowed to reduce the off-policy reinforcement learning problem to the supervised learning task. We tested several combinations of loss functions and output transforming functions. Distillation of knowledge requires a massive experience replay buffer. Several solutions to the problems of optimizing the size of the experience replay buffer are suggested. The first approach is based on the use of a subset of the whole buffer; the second approach uses the autoencoder as a tool to convert states to the latent space. Although our methods can be applied to a wide range of problems, we use Atari games as a testing environment to demonstrate the methods. 

The article provides a brief analysis of the key challenges in the development of mechatronic-modular robots with adaptive kinematic structure, which operational reconfiguration is performed in an automatic mode, depending on the characteristics of the problem and conditions for its solution. It is shown that among the many problems in the development of such robots, one of the most important is in the provision of external sensory functions. The original approach is proposed to create promising samples of the new generation reconfigurable robotic systems with expanded capabilities in sensory, locomotor and other functions based on a complex integration of the mechatronic-modular structures of limited functionality. It is claimed that practical implementation of the proposed approach allows for various ways of its physical implementation and essentially boils down to formation of the multiagent system that combines interacting mechatronic-modular structures, whose movement must be synchronized, in a single composition. As an example, two fundamentally different equipment options of the wheel-configuration modular robot with mechatronic-modular walking "platform" carrying technical vision and being able to move in the opposite to the wheel rotation direction are considered. The algorithms for automatic formation, reconfiguration and motion control for a wheel-configuration mechatronic-modular robot with walking sensory platform are being developed. The approaches of using bulletin board mechanisms to develop the means of behavior planning and interaction coordination of the autonomous self-reconfigurable robots are being discussed. The results of experimental research and computer simulation are presented, confirming the efficiency of the presented algorithms. 

In the interests of improving the quality of a group aircraft flight planning, the developing algorithms problem formulation for the operational determination of the permissible observation duration for ground objects is formulated. An algorithm for determining the observation duration when servicing the next request to be described in the form of a fuzzy logic procedure is proposed. To implement the algorithm for determining the duration of observation, a specialized expert system has been developed. The input of the expert system receives values that describe the influence of the factors in assessing the priority of servicing the next object. At the output of the expert system, an alternative is formed to continue searching for an object or to stop. A new approach to solving the target distribution problem between aircraft in a group flight is proposed. It is based on the joint use of two dynamic priorities in the minimax algorithm for selecting observation objects and for assigning service aircraft. An original approach to determine the rational aircrafts composition in one flight is proposed. This is done with the help of the queuing theory apparatus, which takes into account the random nature of the dynamic situation. To estimate the required aircrafts composition when servicing the flow of requests, the countermeasures process is described using two nonlinear differential equations (of the Riccati type). A general formula for determining aircrafts composition in one flight has been obtained. Mathematical model of the aircraft survivability loss in the form of the Bernoulli equation is formed. Computer simulation of the aircraft survivability losses in one flight was carried out for three cases: with weak interference, with equal opposing forces, with strong counteraction. 

The principles of operation of a wave solid-state gyroscope are described using the example of a cylindrical metal resonator design, in which oscillations are created and a signal is picked up by means of piezoelectric elements located on the base of the resonator. The mode of positional excitation of primary oscillations and the formation of a feedback signal to ensure the operation of the resonator in the mode of an angular velocity sensor are considered. A block diagram of the conversion sequence of the signal taken from the piezoelectric elements of the axes of nodes and antinodes for the organization of feedbacks is given. Approaches to the construction of control loops for the amplitude and phase of oscillations to create primary oscillations and compensate for the amplitude of oscillations of the node axis are considered. The presence of these circuits makes it possible to ensure the operation of a wave solid-state gyroscope in the mode of an angular velocity sensor. For all areas of signal conversion, a mathematical description of computational processes has been compiled, based on the condition for the implementation of computational procedures by a digital computer. Mathematical dependences necessary for correcting the output signal — ensuring its linearity and stability in the range of operating temperatures and the range of measured angular velocities are compiled. The results of tests confirming the operability of a wave solid-state gyroscope in the mode of an angular velocity sensor are presented. An approach to constructing a resonator model based on the finite element method is considered. An approach to the verification of a finite element model is shown by comparing the results of an experimental study of the characteristics of the resonator with the simulation results.