مجله کنترل
سال دوازدهم شماره 4 (زمستان 1397)

In this paper, design of continuous time predictive controller and solving the resulting differential-algebraic equations are presented using the semi-analytical homotopy perturbation method. At any updating time of the continuous time predictive control algorithm, an optimal open loop control problem must be solved. In order to solve the predictive control problem in continuous time, the problem of optimal control is solved by an indirect method. For this purpose, the necessary and sufficient conditions for optimality are determined by applying the variational calculus and the Pontryagin's minimum principle. A system of differential-algebraic equations with boundary conditions is created. Homotopy perturbation method is proposed to semi-analytically solve this problem. By specifying the control and the state functions, we can obtain easily the control and the state values in every instance of the prediction horizon. The presented method can be used to design of continuous-time predictive controller of linear, nonlinear and time varying systems. To illustrate the reliability and efficiency of the proposed method, some numerical examples with simulation results are presented.

In this paper, the problem of joint tracking and system calcification for a maneuvering target has been investigated. The system classification could improve performance of a tracking algorithm in a majority of applications. For instance, it is very crucial to determine the class of target in caring systems like air traffic control, marine care, and air defense at any time. In contrast to the existing solutions, which consider a separate filter for each class, we propose a single particle filter to estimate the class of target leading to a considerable reduction in computation complexity. Simulation results show that the proposed algorithm can estimate the class of target efficiently

The goal of the Doppler and Bearing Tracking (DBT) as a kind of passive target tracking problem is to estimate the position and velocity of the target using its transmitted signal. The main problem of this kind of target tracking is nonlinearity of the measurement equations. In order to solve this problem, different approaches have been reported in the literature, such as extended Kalman filter. However, bias and dependence on the initial conditions are the key shortcomings of such filters. In this paper, first, a novel technique is proposed to provide an appropriate initial condition for the filter. Then, inspired by the modified covariance extended Kalman filter, a new adaptive extended Kalman filter is introduced. Here, the measurement and the process noise covariances are updated simultaneously for reducing the bias effects. Finally, the performance of the proposed filter is compared with the standard extended Kalman filter, adaptive extended Kalman filter and unscented Kalman filter. Results show the good performance of the proposed filter in the problem under study

In this paper, using a new modeling, an Extended Kalman Filter (EKF) is presented for estimation of attitude (i.e. roll and pitch angles) and gyroscope sensor bias using a tri-axes acceleration and a tri-axes gyroscope. The algorithm is developed for accurate estimation of attitude in dynamic conditions and existence of external body acceleration. The external body acceleration estimation as the main source of attitude estimation error in dynamic conditions is very important in attitude estimation accuracy, but in the literatures, the error of the external body acceleration on attitude estimation has not been studied in different dynamic conditions. The paper deals to estimation of the gyroscope sensor bias in two rotational axes (roll and pitch), accurate attitude estimation in different dynamic conditions and estimation of external body acceleration. The proposed algorithm application for attitude, external body acceleration and gyroscope sensor bias is evaluated by quasi-static and dynamic experimental tests in high acceleration bound.

This paper studies the linear positive discrete-time systems with the uncertainty in the system model. Some parameters of system’s matrix are unknown and only some information about their lower and upper bounds are available. The goal of this paper is design of output feedback control law in the presence of parametric uncertainties. The control law is designed in a way that in addition to asymptotic stability of the closed-loop system; it guarantees the positivity of the closed-loop system which adds some complexity in the process of problem solving. Another case has also been investigated in this paper is design of output feedback control law such that as well as the asymptotic stability and positivity of the closed-loop system; the control signal be also positive. The realization of this constraint is necessary in some positive systems. In this paper, two theorems are given and proved. Moreover, the conditions expressed in these theorems are converted to linear programming formats. Finally, computer simulations are presented to verify the theoretical results.

Angular displacement measuring encoder is one of the most important measuring tools in the automated systems, industries and machine tools. In many of manufacturing and production processes, rotary encoders are used as reliable tools for precise positioning. In this study, a new capacitive-type rotary encoder with un-tethered rotor is designed. The main components are made of printed circuit films. Hence, the encoder can be set up in thin inter spaces. The encoder consists of a receiving stator and a transmitter rotor, respectively containing four-phase and two-phase electrodes. In order to designs an un-tethered rotor; the encoder employs a unique approach. Electrical power is supplied to the transmitter rotor by electrostatic induction without any physical contact. This un-tethered rotor can facilitate sensitive applications that a mechanical disturbance caused by an electric wire can be a problem. In this study the encoder was built and its performance was evaluated. The result of experimental evaluation shows that the encoder has max ±0.09 degree error.