System Modeling and Artificial Neural Network (ANN) Design for Lateral and Longitudinal of F-16

Today, classical control methods are still widely used because of their excellent performance in a working enviroment with noise signals. Besides, they are suitble for functiions of the system : operations to control a machine are more flexible, easy to perform, less unwanted risks occur, the efficiency of controlling a system better. In the early years of the 21st century, traditional algorithms still promote their effects. Besides the traditional control methods, the author has applied more moderm and smarter algorithms such as adjusting Linear Quadratic Gaussian (LQG) to control a system on the ground or a system moving in the air. In the paper, LQG regulator is applied to a flight model to demonstrate its effectiveness in all cases. LQG regulator has not been applied before for this model. Results are as expected by the author for the working enviroment with noise signals affecting the system. Kalman filter used in this paper has shown its usefulness in the problem of dealing with unwanted signals. Simulation is done by Matlab.

A new control method for a rectifier based on a transformer with a rotating magnetic field (TRMF) containing an even number of sections of circular winding (CW), which makes it possible to double the number of ripples of rectified voltage in the supply-voltage period, is considered. It leads to a reduction in the ripple amplitude and improves the quality of rectified voltage. The geometrical similarity of the CW of a TRMF making it possible to implement the new control method is analyzed. Switching algorithms of power switches for the classical and new control methods for any quantity of an even number of CW sections are presented. As an example, a switching queue of power switches for a CW with ten sections is considered. The output-voltage shapes for different depths of control are shown. The calculating procedure of the control characteristic for the classical and new control methods, as well as corresponding diagrams, are presented. For the presented example with ten sections of CW, the efficiency of the new control method was evaluated and the dependences of the voltage-ripple ratio on the RMS of rectified voltage are presented. A comparative analysis of the number of ripples for different quantity of CW sections was conducted, and the harmonic composition of output voltage spectrum for the example under consideration is determined. It has been concluded that, in the case of using the new control method for achieving the desired quality of rectified voltage, the number of CW sections may be halved compared with the classical control method, which will simplify the manufacture of a transformer magnetic conductor, halve the number of pairs of power switches switching the CW taps, and decrease the mass, dimension, and cost factors of a semiconductor switch and the TRFM itself as a whole. Using the new control method makes it possible to make a controlled rectifier with an even number of sections of a CW transformer almost as efficient as using a transformer with an uneven number of CW sections and the classical control method. The new control method ensures the fulfillment of the requirements for the available value of the ripple ratio of the output voltage of a controlled rectifier based on a transformer with a rotating magnetic field set out by State Standard YuST 13109-97 at a lower RMS of rectified voltage.

Automatic design of neural network structures

Mobile robots are extensively used in various terrains to handle situations inaccessible to man. Legged robots in particular are tasked to move in uneven terrain. Hence the primary problem of these robots is locomotion in an autonomous fashion. Autonomy is important as the tasks are plagued with many uncertainties. Some of these tasks include leg movement and coordination, navigation, localisation, and stability during movement, all operating in dynamic and unexplored territories. Classical control and traditional programming methods provide stable and simple solutions in a known environment. The environment that legged robots work in is dynamic and unstructured, and such control methods are not always able to cope with. It is difficult to model the environment to provide the controller with the relevant data and program actions for all possible situations. Hence controllers with abilities to learn and to adapt are needed to solve this problem. Soft computing provides an attractive avenue to deal with these situations. Soft computing methods are based on biological systems and they provide the following features: generalisation, adaptation and learning. As more is realised about the use and properties of soft computing methods, the development of controller is shifting towards using soft computing. They have properties that can be used to improve the stability, adaptability, and generalisation of the controllers. Some of the more popular methods used are fuzzy logic, artificial neural networks, reinforcement learning and genetic algorithms. They are commonly integrated with classical methods to enhance the features of classical controllers and vice versa. Each soft computing method serves a different purpose, with its advantages and disadvantages, and the methods are often used together to complement each other. This chapter provides a survey on the different uses of soft computing methods in the different aspects of legged robotics. We see how soft computing methods and classical techniques compliment each other. Two areas of legged robotics are dealt with - control architecture and the problem of navigation. The Central Pattern Generator (CPG) controller and the behaviour-based controller are two architectures presented in this chapter. Various soft computing techniques are used to implement and improve these two controllers.

Service robots are built and developed for various applications to support humans as companion, caretaker, or domestic support. As the number of elderly people grows, service robots will be in increasing demand. Particularly, one of the main tasks performed by elderly people, and others, is the complex task of cleaning. Therefore, cleaning tasks, such as sweeping floors, washing dishes, and wiping windows, have been developed for the domestic environment using service robots or robot manipulators with several control approaches. This article is primarily focused on control methodology used for cleaning tasks. Specifically, this work mainly discusses classical control and learning-based controlled methods. The classical control approaches, which consist of position control, force control, and impedance control , are commonly used for cleaning purposes in a highly controlled environment. However, classical control methods cannot be generalized for cluttered environment so that learning-based control methods could be an alternative solution. Learning-based control methods for cleaning tasks can encompass three approaches: learning from demonstration (LfD), supervised learning (SL), and reinforcement learning (RL). These control approaches have their own capabilities to generalize the cleaning tasks in the new environment. For example, LfD, which many research groups have used for cleaning tasks, can generate complex cleaning trajectories based on human demonstration. Also, SL can support the prediction of dirt areas and cleaning motion using large number of data set. Finally, RL can learn cleaning actions and interact with the new environment by the robot itself. In this context, this article aims to provide a general overview of robotic cleaning tasks based on different types of control methods using manipulator. It also suggest a description of the future directions of cleaning tasks based on the evaluation of the control approaches.

This paper presents the design of radial basis function geometric bioinspired networks and their applications. Until now, the design of neural networks has been inspired by the biological models of neural networks but mostly using vector calculus and linear algebra. However, these designs have never shown the role of geometric computing. The question is how biological neural networks handle complex geometric representations involving Lie group operations like rotations. Even though the actual artificial neural networks are biologically inspired, they are just models which cannot reproduce a plausible biological process. Until now researchers have not shown how, using these models, one can incorporate them into the processing of geometric computing. Here, for the first time in the artificial neural networks domain, we address this issue by designing a kind of geometric RBF using the geometric algebra framework. As a result, using our artificial networks, we show how geometric computing can be carried out by the artificial neural networks. Such geometric neural networks have a great potential in robot vision. This is the most important aspect of this contribution to propose artificial geometric neural networks for challenging tasks in perception and action. In our experimental analysis, we show the applicability of our geometric designs, and present interesting experiments using 2-D data of real images and 3-D screw axis data. In general, our models should be used to process different types of inputs, such as visual cues, touch (texture, elasticity, temperature), taste, and sound. One important task of a perception-action system is to fuse a variety of cues coming from the environment and relate them via a sensor-motor manifold with motor modules to carry out diverse reasoned actions.

Artificial Neural Networks (ANN) and Deep Learning (DL) are used to solve complex problems including image recognition, speech recognition and have applications in new technologies for autonomous driving, facial recognition, detecting cancers from imaging samples among others. Various design considerations are involved in the design, training, and testing of Artificial Neural Networks (ANN). These include the design of the input/output layer, the structure and number of hidden layers, the data/data-structures of variables, the transformative functions embedded in the network, the optimizers being considered, the learning rate and its systematic adjustment, the prudent usage of dropout, the parallelism-related batch-size, the number of epochs, the adaptive logic for systematically changing the network for better fit, etc. While all these methods and techniques are sensible and relevant, there lacks an overarching framework for the needed design. This paper considers the design of an ANN from an Axiomatic Design (AD) perspective that parallels the biological inspiration for ANN’s in the first place, i.e., the brain. The axiomatic design approach is used for explicating and extricating the form, function, and adaptive evolution of the underlying network.

The intensive monitoring of air pollutants has led to the acquisition of vast quantities of data. Traditional quality control methods based on existing knowledge may be inefficient because of our limited understanding regarding the interaction of human activities and stochastic environmental factors. Moreover, traditional methods for outlier detection may be misleading because of the existence of valid outliers and invalid inliers. In this research, artificial neural networks (ANNs) are developed to identify instrument failure based on current and historical observations. Two structures, i.e., multilayer perceptrons and recurrent networks, are trained using 50,000 hourly data points labeled by human reviewers. The most conservative model identified 57.5% of the invalid sulfur compound observations and 44.9% of the invalid nitrogen compound observations. By setting a more liberal threshold, these values increased to 76.0% and 79.7%, respectively. Except for SO2, the ANNs outperformed the traditional methods for data quality control, as demonstrated with a plausibility test, a test of temporal consistency and a residential analysis. Compared with the test of temporal consistency, which was the most effective traditional method studied, the true positive rates of the ANNs were 19.4% to 29.5% higher for all pollutants except SO2, given the same false positive rates. The results indicate the effectiveness of ANNs for data quality control even without supplementary information. Methods for performance improvement are discussed.

In recent years, the Biped Robot is more and more self-determining and time-sensitive, so the stability has become a very important question. But the traditional control methods can not meet it. To solve this question, Artificial Neural Network (ANN) has been brought up. Instead of most traditional control methods, Artificial Neural Network is applied widely to control the Biped Robot to walk accurately and stably. In this work, we design a control system of the Biped Robot with GA-ANN (Artificial Neural Network based on Genetic Algorithm). The GA-ANN control system adjusts the weights by the robot's Zero Moment Point (ZMP), tracks the robot's nonlinear kinetic system and keeps the robot step stably. Experiments show the stability improvement of robot using proposed algorithm.

Design of neural networks using genetic algorithm for the permeability estimation of the reservoir

The use of modern and traditional methods of fertility control in Bangladesh: A multivariate analysis

High-speed brushless dc motor (BLDCM) has been widely used in many applications, but the rotor eddy current loss has always been a key constraint by traditional PWM control methods because of the serious distortion of stator current. Although this loss is only a small portion in the overall power loss, it still may expose the magnets to high risk of overheating and demagnetization, especially for the inner-rotor motors with weak cooling condition and motors operated at high-speed region. This paper presents a novel control method, asymmetrical dual-pulse (ASDP) variable conduction period control method, for high-speed permanent magnet BLDC motors to reduce the rotor eddy current loss. Comparative studies of stator current waveforms, harmonic components, and rotor eddy current loss have been carried out between the ASDP method and several traditional control methods. The FEM simulation and THD analysis results confirm that the ASDP method can effectively eliminate the even harmonic and weaken the odd harmonic components in stator current, so that the rotor eddy current loss can be significantly reduced.

Design of an artificial bionic neural network to control fish-robot's locomotion

Because of the disadvantages of the classical control methods, such as its fixed parameters, so the control effect of the classical control methods is greatly limited, and the Biological Intelligence Algorithm has provided a new way to break the bottleneck of classical control methods because of its adaptive and learning mechanism. With the improvement of the theory of reinforcement learning and deep learning, these theories have greatly improved the performance of the Biological Intelligence Algorithm. This paper summarizes seven kinds of intelligent algorithms which is used in intelligence control commonly, and emphatically analyzes the application examples of the combination of classical automatic control method and intelligent algorithms, especially reinforcement learning. The development status and future development trend of intelligent control based on reinforcement learning, deep learning and Brain-inspired Intelligence Technology in recent years are discussed for the first time in this paper. The purpose of this paper is to emphasize a new idea of combining intelligent algorithm with classical control method, and provide new ideas and practical examples for the research of intelligent control.

Integration of wind energy conversion systems (WECS) to the electric grid poses power quality problems due to the intermittent nature of wind flow. Power quality improvement has been addressed using conventional approaches for the adjustment of wind turbine’s tip speed ratio (TSR). However, low rotor speed and mechanical sensor failure are linked to the traditional classical control (TCC) methods. Therefore, this study employed an artificial neural network (ANN) technique in adjusting turbine’s TSR for power quality enhancement. Data on windspeed and turbine hub-speed were obtained from Shagari Wind Farm in Sokoto, Nigeria and a mathematical model of ANN-based TSR controller was develop. Using MATLAB/Simulink platform and Levenberg-Marquardt algorithm, a simulation model of the proposed method was developed. Impact of the TSR controller on the turbine blades was investigated and evaluated together with the quality of the microgrid’s output power, using mean square error. With implementation of the method on a 2.5 KW rated WECS and with TSR variations from 0 radians to 25 radians, the mechanical power output using the TCC method becomes stable at 240W level, while for the proposed method the stability level was achieved at 140W; indicating 58.33% improvement in the stability responses. In comparison with the TCC approach to TSR adjustment, the technique proposed in this study offers a better performance. Therefore, it is recommended for the improvement of power quality on grid-connected WECS.