Performance and improvement of flight efficiency at various velocities for flight systems, in particular, rotorcrafts, with specific complexities in motion and its nonlinear equations are always of particular interest to researchers in the aerial and control domains. In this research, a new control algorithm is addressed based on the complete nonlinear Unmanned Rotorcraft (UR) model and its four main inputs. Exploiting state feedback and Polytopic Linear Parameter Varying (PLPV) modeling and using Linear Matrix Inequality (LMI), the velocity control problem is investigated. The trim points of the system are produced under different velocity control conditions. State feedback control gain matrix which plays a main role in producing the ultimate control signal, is computed by solving a set of LMIs under various conditions. Finally, instead of using a Nonlinear model, a Polytopic model is used for controller synthesis. With this goal, different scenarios for the proposed flight velocity control (in different dynamic ranges, minimum velocity to maximum velocity) are implemented. The simulation results demonstrate a very good performance of the proposed controller in the basis of PLPV modelling. It can be concluded that the proposed manner is useful to overcome the disruptions imposed on the flight system due to the changes in the equilibrium points and the uncertainties of the parameters and/or possible errors due to the unwanted possibilities in the system. پرونده مقاله

This paper presents a new method for modeling and Attitude Control of Autonomous Helicopters (A.H.) based on a polytopic linear parameter varying approach using parameter set mapping with the Principal Component Analysis (PCA). The polytopic LPV model is extracted based on angular velocities and Euler angles, that is influenced by flopping angles, by generating a set of data over the different trim points. Because of the high volume of trim data, parameter set mapping based on (PCA) is used to reduce the parameter set dimension. State feedback control law is proposed to stabilize the system by introducing a Linear Matrix Inequality (LMI) set over the vertices models. The proposed controller is performed for an Autonomous Helicopter in different scenarios. All the scenarios are investigated with the PCA algorithm as a technique for reducing the computational volume and increasing the speed of solving the LMI set. The simulation results of implementing the planned controller on the nonlinear model of an autonomous helicopter in different scenarios show the effectiveness of the proposed scheme. پرونده مقاله

In this investigation, automatic face recognition algorithms are discussed. For this purpose, a combination of learning algorithms with supervision are realized; in this way, the classification is first designed by the fuzzy-based support vector machine and then the AdaBoost meta-algorithm is applied to the designed classification to reach more accuracy and overfitting control. In the research proposed here, in order to address the effects of asymmetric classes, the adaptive coefficients are employed. In addition, to reduce the data size, the principal components analysis is also applied to the raw data. It is to note that the proposed approach is carried out in a set of images extracted from Yale University data set and its accuracy of the proposed one is verified. پرونده مقاله

The length between the current point in the degradation process and the time of reaching the failure threshold, or the remaining usable life (RUL) prediction of systems, is of the greatest priority in the industry. More accurate estimation is useful for maintenance decisions as it helps to avoid catastrophic breakdowns and may also assist in reducing additional costs. Deep learning approaches have made impressive advancements in this field in recent years by becoming widely attractive and employed. However, most deep learning approaches don’t fully consider the information implications of sensors adaptively. To overcome this problem, a novel adaptive hybrid model that combines a convolutional neural network (CNN) and gated recurrent unit (GRU) is introduced in this work. The RUL estimation is based on the best practical option of sequence data through CNN-GRU. In the first step, optimal sensor selection is applied to the dataset to collect the most useful sensors. Then, the input data is transformed into a predefined range of values using standard and min-max scalars; in the next step, the normalized data is fed into the CNN-GRU model with an adaptive activation function for deep feature extraction, training, and RUL prediction. Utilizing CNN to extract features from the multivariate input data automatically, the features are then fed into the GRU layer to train the model for RUL prediction. To test the effectiveness of this framework, the suggested methodology is applied to the NASA Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) dataset. The findings demonstrate that CNN-GRU is capable of accurate RUL prediction. In addition, CNN-GRU outperforms CNN-LSTM and CNN-RNN in terms of computation efficiency and accuracy. پرونده مقاله