parameter estimation

In order to perform attitude control ground experiments, it is crucial to ensure the center of mass of the platform is coincident with the center of rotation as accurate as possible. In this paper, a linearized dynamic model of the imbalanced attitude platform is derived and a linear controller is methodically designed to stabilize the system by mass relocation. The stability conditions of the system under the proposed controller are derived. The balancing procedure starts with a parameter estimation method to estimate the center of mass offsets. Next, the PD controller is applied to align the platform’s horizontal plane with the local horizontal level. Finally, the imbalance in vertical axis can be estimated and compensated to complete the automatic mass balancing. Actuator limitations and nonlinear equations of motion are implemented in numerical simulations and the results demonstrate the effectiveness of the proposed method in significantly reducing the residual imbalance torque on the simulated platform.

System Identification is used to derive the dynamic and mathematical model of a system using experimental data. If the case study system is an aircraft, this process is called aircraft system identification, which is used to estimate the stability and control derivatives of the aircraft. The least squares method is one of the methods used in the field of time and frequency to estimate the parameters of the aircraft. In this paper, system identification and estimation of F-16 aerodynamic coefficients are performed using the least squares method and the equation error approach. Hence the main focus is on the formulation of the aerodynamic model for estimating unknown coefficients using the least squares method. In this regard, nonlinear 6-DOF of the F-16 aircraft was simulated and using its results, system identification and estimation of transfer function, force coefficients and aerodynamic moments of the aircraft lateral-directional channel using regression of least squares and the results are analyzed. The results show that the modeling and estimation of the coefficients is compatible with the experimental data and can be suitable for further studies, including control objectives.

Development of low-cost small satellites has been at the center of attention in recent years. Concurrent Orbit and Attitude Estimation (COAE) requires fewer sensors onboard and subsequently results in some cost reductions. In this regard, the present paper has focused on addressing the importance of COAE utilizing temperature rate on satellite surfaces. To this end, the thermal model for a low Earth orbiting satellite is introduced first. A three-axis stabilized spacecraft is assumed equipped with small measurement plates that are isolated from each other and from the internal heat sources of the satellite. As the Sun and the Earth are the significant sources of radiation for a near Earth space system, the view factor is the key parameter for observability of the orbital elements, while the Sun radiation is responsible for the attitude observability. The Earth albedo factor is a major uncertain parameter required for the thermal analysis of low Earth orbiting satellites. This parameter is greatly dependent on the Earth’s local terrain and climatic conditions such as instantaneous cloud coverage. To address the problem of albedo factor uncertainty, it is estimated simultaneously with the attitude and orbit of the satellite. NASA's CERES project provides satellite-based observations of the Earth’s radiation budget and clouds over almost 18 years. In this paper, CERES data tables for the Earth’s thermal flux and albedo factor have been used to produce more realistic measurement data. The nonlinear filter of Unscented Kalman Filter (UKF) is also exploited for the state estimation. Lack of sun radiation during the satellite’s eclipse intervals results in the loss of orbit and attitude observability. The performance and viability of the proposed COAE algorithm are verified by Monte Carlo simulations. Moreover, a sensitivity analysis is conducted within a wide range of semi major axes, eccentricities, and inclinations. The results demonstrate the high sensitivity of the algorithm to the orbit altitude and the sun rays direction.

The sonar system is one of the tools for detecting moving marine vessels such as high speed crafts. In marine vessels detection, due to the high speed of movement, the received signal is combined with ambient noise and affected by multi path, therefore it is necessary for the system to know the environmental parameters. The underwater environment is a complex environment that is affected by various physical factors and due to the non-stationary and nonlinear nature of the phenomena, it is very difficult or impossible to study all the parameters affecting the sound wave. But for the analysis of acoustic phenomena, such as how the wave propagates, how the signal attenuates in dealing with different surfaces and channel modeling, according to theories of physics and mathematics, different statistical analyzes were presented. One of these models is the K probability distribution model, which is used to study the underwater acoustic reverberation. In this paper, a scenario of underwater channel reverberation are simulated and a resonant signal with k distribution are generated. Then, based on the fitting of the curve of its cumulative distribution function, the k distribution parameter is estimated using the optimization algorithm. The proposed method is compared with the some methods such as maximum likelihood method, the method of moment and the zlogz method, which the results of the analysis show the superiority of the proposed method over other methods.

Tomography is a non-invasive method that is used to visualize internal structure of an object. Electrical tomography is a method that recognizes the geometry of an object by applying voltage or current and using an image reconstruction algorithm. In spite of some advantages such as simplicity and low cost, reconstructed images by this method have low resolution and low quality. In addition to the system hardware, the reconstruction algorithms also reduces the quality of images. In this research, Levenberg – Marquardt Algorithm which is used to solve inverse heat transfer problems, is employed for image reconstruction in electrical tomography. In this technique, image reconstruction is perfomed by changing the thermal conductivity of material. Three different test cases are selected to investigate the capability of proposed algorithm in estimation of the unknown geometry and the results show that the Levenberg – Marquardt algorithm has ability to estimate unknown shapes. Sensitivity analysis is also performed in order to examine the effect of noise in detection of the unknown geometry. The results showe that with increasing the noise, the error created in the estimation of shape increases, but the estimated shape has good agreement with original geometry. Also, results of choosing different combinations of active sensors for applying thermal flux show that in addition to effectiveness of active sensors in increasing the accuracy of shape estimation, the unknown geometry can be estimated with two measurements.

In this paper, a new model-based fault detection method for an agile supersonic flight vehicle is presented. A nonlinear model, controlled by a classical closed loop controller and proportional navigation guidance in interception scenario, describes the behavior of the vehicle. The proposed FDD method uses Inertial measurement unit (IMU) data and nonlinear dynamic model of the vehicle to inform fins damage to the controller before leading to an undesired performance or mission failure. Broken, burnt or not opening of control surfaces causes a drastic change in aerodynamic coefficients and thus dynamic model. Therefore, in addition to the changes in the control forces and moments, system dynamics will change too, leading to the failure detection process being encountered with difficulty. To this purpose, an equivalent aerodynamic model is proposed to represent the dynamics of the vehicle, and the health of each fin is monitored by the value of a parameter which is estimated using a nonlinear filter. To identify model changes after fin failure, the aerodynamics of body and control surfaces are modeled separately. A percentage of the fins which have the ability to generate aerodynamic force is modeled by a parameter and these parameters are estimated over time, using a nonlinear estimator. Parameters estimation through the filtering approach is an indirect procedure, consisting of transforming the problem into a state estimation problem. The value of these parameters will be between 0 and 1. Where value 1 corresponds to the health of the fin and value denotes whether the fin has not opened or has been destroyed completely. The proposed method detects and isolates fins damage in a few seconds with good accuracy. Simulation results show that the new failure diagnosis algorithm estimated fins health percentage with good accuracy. After estimating the parameters, they can be sent to the controller and the controller has changed control signal.

A major problem in design of a controller is that some of the parameters of the governing dynamic equation might be unknown. The purpose of the present study is to provide a method for identification of the unknown parameters of the dynamic equations of the two-axis gimbal system and then applying a suitable controller to it. Among the uncertain parameters which may exist in the dynamic equations of a two-axis gimbal, the components of the inertia matrix, the coriolis matrix, and the friction matrix can be named. In this research, in order to estimate the unknown parameters of the system, the Gauss-Newton method which is a gradient-based inverse technique is used. Regarding the severe sensitivity of inverse problems to measurement errors, by using a proper smoothing technique, this undesirable effect is reduced. In the present work, for the first time, an accurate technique for computation of sensitivity coefficients of the dynamic equations of the gimbal is presented. After identification of the unknown parameters, the control of the two-axis gimbal is also performed by sliding mode control. The applied moments are designed by sliding mode control to stabilize the line of sight. Also, proof of stability is presented for the sliding mode control. Simulation of the laboratory tests in the identification section and also control results are obtained by modeling of the system in MATLAB.

Activated carbon (AC) is a kind of carbon that because of having tiny and low-volume cavities provides a good contact surface for adsorption and chemical reactions. AC has various applications in industries like filtration, water purification, refrigeration, agriculture, medicine industries and so on. Recognizing thermal properties of AC is valuable to provide an appropriate design of its applications especially in adsorption refrigeration cycles. In such designs the most important parameters of AC bulk is its thermal conductivity and thermal contact conductance resistance with wall. In this study, a test setup is prepared to conduct experimental investigation and estimate required parameters for granular AC. Granular AC with four different densities was packed in a steel closed cylinder and thermal loading was applied on its outside wall. By measuring transient temperature at the center of the cylinder and using inverse thermal methods, thermal conductivity and contact conductance with steel wall were estimated. Results show increase in thermal conductivity and contact conductance between carbon and steel wall as density increases.

Appropriate control of the vehicle longitudinal dynamic requires accurate and real-time knowledge of the parameters affecting the dynamics of the vehicle. The main contribution of this paper is in developing an estimator called short-time linear quadratic form (STLQF) which is a new parameter estimation technique to simultaneously estimate the time-varying parameters that affect on vehicle longitudinal dynamics with a low latency. These parameters are the vehicle mass, time varying road slope angle and overall aerodynamic drag coefficient of the vehicle in real time. Next, the efficiency of STLQF algorithm is shown by comparing the results of STLQF technique in an experimental test against the recursive least squares (RLS) parameter estimation method. The results of implementing STLQF parameter estimation algorithm, showes the better performance of STLQF estimator compared with RLS algorithm. In the end, the STLQF estimator is simulated in hardware in the loop configuration using an AVR microcontroller in order to rapid prototype of the STLQF method by minimum cost and high accuracy in the least possible time.

In this paper, after presenting a brief introduction about Amirkabir University of Technology’s attitude simulator, governing equations are obtained. Viscous friction model is chosen to model the existing friction in the bearings of the attitude simulator. The main purpose of this paper is to design an adaptive attitude control algorithm for the attitude simulator in order to control it in the desired path and estimate the coefficients of friction due to simulator’s bearings. In order to prevent singularity in simulations, Rodriguez parameters are used for kinematics representation. Firstly, the governing equations are transformed into a robotic form. The moments of inertia and coefficients of viscous friction model are assumed as uncertainties. Then, by introducing a Lyapunov function, the stability of the system is checked and the parameters are estimated. The adaptation law is obtained by the Lyapunov function and the stability of the system is then proved. In order to demonstrate the efficiency of this adaptive control algorithm, a nonlinear Lyapunov-based attitude control algorithm is designed and compared to the adaptive controller. The simulations are done in Matlab software package and the parameters of the moment of inertia matrix and coefficients of viscous friction model are estimated by the adaptation law of the controller. During the simulation, the rotational velocity of the reaction wheels are obtained and it is shown that this attitude control algorithm is implementable on the attitude simulator.

In this paper, a numerical algorithm based inverse method is used to estimate effective diffusion coefficient by using experimental tracer distribution. The Algorithm uses factitious experimental data which are produced by adding noise to numerical data obtained from direct problem. A comprehensive model (Diffusion-Convection-Reaction) is used to derive PET tracer distribution in tumor tissue with microvasculature network. This model was used because of considering all transport phenomena in tissue. In this work to achieve accurate distribution of tracer in tumor tissue, convection diffusion reaction equation which is a PDE is implemented. The proposed tracer in this work is Fluorodeoxyglucose (18F). Solution of inverse problem for estimating effective Diffusion Coefficient is based on minimization of least squares norm. In this work Levenberg-Marquardt technique is applied. Solution of parameter estimation problem require calculation of sensitivity matrix which elements are sensitivity coefficients. Sensitivity coefficients shows differentiation of Tracer concentration with respect to Effective Diffusion coefficient variation is calculated using first derivation of concentration equation. The equations of concentration distribution and sensitivity coefficients are solved using Finite volume method. The results show that the numerical algorithm is able to estimate the effective diffusion coefficient in tissue.

In this paper, the diffusion coefficient in a normal tissue and tumor are to be estimated by the method of inverse problems. At the beginning, distribution of drug (with the assumption of uniform and isentropic diffusion coefficient) in the tissue is considered as the direct problem. In the direct problem, the governing equation is the convection–diffusion, which is the generalized form of fick’s law. Here, a source and a sink are defined; the source as the rate of solute transport per unit volume from blood vessels into the interstitial space and the sink as the rate of solute transport per unit volume from the interstitial space into lymph vessels are added to this equation. To solve the direct problem, the finite difference method has been considered. Additionally, the diffusion coefficient of a normal tissue and tumor will be approximated by parameter estimation method of Levenberg-Marquardt. This method is based on minimizing the sum of squared errors which in the present study, considered error is the difference of the estimated concentration and the concentration measured by medical images (simulated numerically). Finally, the results obtained by Levenberg-Marquardt method have provided an acceptable estimation of diffusion coefficient in normal tissue and tumor.

Cell injection system in medicine used to inject the materials into the cells. The injection system consists of Injector and rotating plate. The controller sets height, position and orientation of the rotating plate. The proposal of this article is to replace SCARA robot injection tool and it included ability in desired position tracking and applied to time-varying force. In recent articles the control system applies to the rotating plate of Cells and this method can cause the damaging risk. The proposed method is fixed plate and to increase the success rate, the robot had been controlled. The parameters of environmental models are estimated by nonlinear proposed models and by using the recursive method, the minimum of squares errors will be optimal. The voltage strategy can control robot actuators. This method is simpler and free from the manipulator dynamics. In all recent studies, the impedance control is based on the torque control method and the proposed method of this article is applying the impedance control using voltage control. The robust adaptive impedance controller is designed in the presence of uncertainties. The simulation's results demonstrate desired performance of the proposed method

In this paper, a trajectory tracking control strategy for a quadrotor flying robot is developed. At first, dynamic model is obtained by lagrange-euler approach. Then, control structure, consisting of a model-based predictive controller, has been used based on state space error to track transitional movements for reference trajectory and also robust nonlinear H∞ control is applied for stabilizing the rotational movements and reject the external disturbance. In both controllers the integral of the position error is considered, allowing the achievement of a null steady-state error when sustained disturbances are acting on the system. The external disturbances is considered as aerodynamic torques. If uncertainties increase, the designed control system will be unable to track and stabilizing perform properly and completely. So finally, in order to eliminate the effects of parameter uncertainties the recursive least squares is used for estimating mass and moment inertia parameters which are linear and it is applied to the control system. Simulation results show that by using estimation of system parameters, the proposed control system has a promising performance in terms of stabilization and position tracking even in the presence of external disturbance and parametric uncertainties.

The Unscented Kalman filter (UKF) is the popular approach to estimate the recursive parameter of nonlinear dynamical system corrupted with Gaussian and white noises. Also, it has been applied to train the weights of the multi-layered neural network (MNN) models. The Group method of data handling (GMDH)-type neural network is one of the most widely used neural networks which has high capacity in modeling of the complex data. In many researches, different approaches are used in training of neural networks in terms of associated weights or coefficients, such as singular value decomposition, and genetic algorithms. In this paper, the unscented Kalman filter is used to train the parameters of GMDH-type neural network when the experimental data are deterministic. The effectiveness of GMDH-type neural network with UKF algorithm is demonstrated by the modeling of the using a table of the multi input-single output experimental data. The simulation result shows that the UKF-based GMDH algorithm perform well in modeling of nonlinear systems in comparison with the results of using traditional GMDH-type neural network and is more robust against the model and measurement uncertainty.

In this paper، mass flow rate and location of leakage in natural gas pipeline has been estimated simultaneously using inverse analysis. For doing so، at first natural gas transient flow in pipeline has been simulated numerically; this simulation is named direct problem. In the direct problem، it is assumed that the mass flow rate and location of leakage is definite and the governing equations are inhomogeneous well-known Euler equations. In these equations، the leakage effect has been considered as a source term. Steger–Warming flux splitting method has been used for numerical analysis of these equations. Then the location and mass flow rate of gas leakage of pipeline have been estimated simultaneously using Levenberg-Marquardt method for parameter estimation. This method is an iterative algorithm and based on minimizing the sum of the squares of the errors which are difference between pressures computed by the direct problem and pressures measured by pressure gauges. The results of the direct problem have good agreement with Mac–Cormack method and characteristics method of specified time intervals. The results of the inverse analysis demonstrate that Levenberg-Marquardt algorithm is stable and efficient enough to estimate simultaneously the mass flow rate and location of leakage in natural gas pipeline.