In this paper, the estimation problem for the sensors network in anisotropy-based theory is considered. The non-centered external disturbance belongs to sequence of random vectors with bounded anisotropy of extended vector. Sensors of the network are assumed to be non-ideal with given probability of correct operation. The communication scheme for networks is described by directed graph. The performance criterion is considered as anisotropic norm for input-to-estimation error of the combined system. The estimation problem is reduced to convex optimisation problem.

The time-optimal control problem for two nonsynchronous oscillators with limited scalar control is considered in cases where the control has two, three and four switches. Algebraic equations for determining switching moments are found, and borders between classes of optimal solutions are analytically obtained on the phase plane of the first oscillator.

This paper describes a control system for two-wheeled mobile robots capable of moving in rooms in an autonomous mode without using satellite navigation signals. The movement takes place along a given route, maneuvering is carried out using an electronic differential that controls the speeds of rotation of both wheels. In this case, the control of the control parameters of the movement is carried out not by odometers or rotation sensors, but by microwave Doppler sensors. This enables a direct measurement of the linear speeds of the wheels. The algorithm of operation of the proposed control system when performing maneuvers and moving along the selected trajectory is described. An increase in the accuracy of movement along the route is noted in comparison with traditional control methods.

В статье представлен подход к созданию интеллектуальной системы прогнозирования в реальном времени состояний технологического процесса на основе анализа видеоряда изображений, получаемых в результате потоковой съемки видеокамерами, установленными на фурмах доменной печи. Предложены алгоритмы распознавания видеообразов фурменных очагов, а также сценарного прогнозирования эволюции технологических ситуаций. Приводится историческая справка о развитии методов автоматического управления доменным процессом, в частности, об использовании методов искусственного интеллекта. Исследование, результаты которого представлены в настоящей статье, направлены на получение возможности оперативно осуществлять анализ производственной ситуации и прогнозирование ее эволюции в ходе реального функционирования доменного процесса, что позволит в автоматическом либо автоматизированном режиме своевременно принимать решения по корректировке управления. На основе выявляемых закономерностей в изменении видеоданных и посредством разработанного авторами алгоритма анализа и прогнозирования динамики технологического процесса предложен метод раннего обнаружения тенденции возникновения определенных ситуаций на фурмах, в том числе - приводящих к дестабилизации технологического процесса в доменной печи...

One of the main disadvantages of artificial neural networks is slow learning associated with the need to calculate a large number of coefficients. The article shows that learning can be substantially accelerated. Acceleration is achieved by a sharp reduction in the number of training patterns. In addition, the inverse pattern method was used both for the formation of features and for the subsequent recognition of objects; this has made it possible to dispense with coefficients, which significantly reduces the amount of calculations. In instantaneous learning, as in deep learning, features are generated automatically. Computational experiments have shown the invariance of the method with respect to not only scaling and rotations but also large deformations of objects to be recognized