scholarly journals Neural Networks for Self-tuning Control Systems

10.14311/514 ◽  
2004 ◽  
Vol 44 (1) ◽  
Author(s):  
A. Noriega Ponce ◽  
A. Aguado Behar ◽  
A. Ordaz Hernández ◽  
V. Rauch Sitar
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Control Systems ◽  
Control Algorithm ◽  
Set Point ◽  
Tuning Mechanism ◽  
Self Tuning ◽  
The Moment ◽  
Weighting Coefficients ◽  
Previous Training

In this paper, we presented a self-tuning control algorithm based on a three layers perceptron type neural network. The proposed algorithm is advantageous in the sense that practically a previous training of the net is not required and some changes in the set-point are generally enough to adjust the learning coefficient. Optionally, it is possible to introduce a self-tuning mechanism of the learning coefficient although by the moment it is not possible to give final conclusions about this possibility. The proposed algorithm has the special feature that the regulation error instead of the net output error is retropropagated for the weighting coefficients modifications. 

The synthesis method of regulators for multichannel systems using neural networks

2020 ◽  
pp. 111-118
Author(s):  
Александр Александрович Воевода ◽  
Дмитрий Олегович Романников
Keyword(s):  
Neural Networks ◽  
Reinforcement Learning ◽  
Automatic Control ◽  
Control Systems ◽  
Synthesis Method ◽  
Neural Net ◽  
Initial State ◽  
Set Point ◽  
Automatic Control Systems ◽  
Policy Gradient

Синтез регуляторов для многоканальных систем - актуальная и сложная задача. Одним из возможных способов синтеза является применение нейронных сетей. Нейронный регулятор либо обучают на предварительно рассчитанных данных, либо используют для настройки параметров ПИД-регулятора из начального устойчивого положения замкнутой системы. Предложено использовать нейронные сети для регулирования двухканального объекта, при этом обучение будет выполняться из неустойчивого (произвольного) начального положения с применением методов обучения нейронных сетей с подкреплением. Предложена структура нейронной сети и замкнутой системы, в которой уставка задается при помощи входного параметра нейронной сети регулятора The problem for synthesis of automatic control systems is hard, especially for multichannel objects. One of the approaches is the use of neural networks. For the approaches that are based on the use of reinforcement learning, there is an additional issue - supporting of range of values for the set points. The method of synthesis of automatic control systems using neural networks and the process of its learning with reinforcement learning that allows neural networks learning for supporting regulation is proposed in the predefined range of set points. The main steps of the method are 1) to form a neural net input as a state of the object and system set point; 2) to perform modelling of the system with a set of randomly generated set points from the desired range; 3) to perform a one-step of the learning using the Deterministic Policy Gradient method. The originality of the proposed method is that, in contrast to existing methods of using a neural network to synthesize a controller, the proposed method allows training a controller from an unstable initial state in a closed system and set of a range of set points. The method was applied to the problem of stabilizing the outputs of a two-channel object, for which stabilization both outputs and the first near the input set point is required

scholarly journals Braking Control for Improving Ride Comfort

2018 ◽  
Vol 166 ◽  
pp. 02002 ◽  
Author(s):  
Jonghyup Lee ◽  
Seibum Choi
Keyword(s):  
Control Systems ◽  
Control Algorithm ◽  
Ride Comfort ◽  
Low Risk ◽  
Vehicle Safety ◽  
Phase Lag ◽  
Vehicle Performance ◽  
High Speeds ◽  
Braking Control ◽  
The Moment

While many vehicle control systems focus on vehicle safety and vehicle performance at high speeds, most driving conditions are very low risk situations. In such a driving situation, the ride comfort of the vehicle is the most important performance index of the vehicle. Electro mechanical brake (EMB) and other brake-by-wire (BBW) systems have been actively researched. As a result, braking actuators in vehicles are more freely controllable, and research on improving ride comfort is also possible. In this study, we develop a control algorithm that dramatically improves ride comfort in low risk braking situations. A method for minimizing the inconvenience of a passenger due to a suddenly changing acceleration at the moment when the vehicle is stopped is presented. For this purpose, an acceleration trajectory is generated that minimizes the discomfort index defined by the change in acceleration, jerk. A controller is also designed to track this trajectory. The algorithm that updates the trajectory is designed considering the error due to the phase lag occurring in the controller and the plant. In order to verify the performance of this controller, simulation verification is completed using a car simulator, Carsim. As a result, it is confirmed that the ride comfort is dramatically improved.

Artificial neural networks in the problem of determining the position and size of the solar disk in wide-angle photographs of the sky

2021 ◽  
Author(s):  
Mikhail Borisov ◽  
Mikhail Krinitskiy
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Solar Disk ◽  
Weather Conditions ◽  
Wide Angle ◽  
Know How ◽  
The Neural Network ◽  
Artificial Neural ◽  
The Moment

<p>Total cloud score is a characteristic of weather conditions. At the moment, there are algorithms that automatically calculate cloudiness based on a photograph of the sky These algorithms do not know how to find the solar disk, so their work is not absolutely accurate.</p><p>To create an algorithm that solves this data, the data used, obtained as a result of sea research voyages, is used, which is marked up for training the neural network.</p><p>As a result of the work, an algorithm was obtained based on neural networks, based on a photograph of the sky, in order to determine the size and position of the solar disk, other algorithms can be used to work with images of the visible hemisphere of the sky.</p>

scholarly journals Verification of Closed-loop Systems with Neural Network Controllers

10.29007/btv1 ◽  
2019 ◽  
Author(s):  
Diego Manzanas Lopez ◽  
Patrick Musau ◽  
Hoang-Dung Tran ◽  
Taylor T. Johnson
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Control Systems ◽  
Closed Loop ◽  
Closed Loop Control ◽  
Feed Forward ◽  
Loop Control ◽  
Closed Loop Systems ◽  
Feed Forward Neural Networks ◽  
Neural Network Controllers

This benchmark suite presents a detailed description of a series of closed-loop control systems with artificial neural network controllers. In many applications, feed-forward neural networks are heavily involved in the implementation of controllers by learning and representing control laws through several methods such as model predictive control (MPC) and reinforcement learning (RL). The type of networks that we consider in this manuscript are feed-forward neural networks consisting of multiple hidden layers with ReLU activation functions and a linear activation function in the output layer. While neural network con- trollers have been able to achieve desirable performance in many contexts, they also present a unique challenge in that it is difficult to provide any guarantees about the correctness of their behavior or reason about the stability a system that employs their use. Thus, from a controls perspective, it is necessary to verify them in conjunction with their corresponding plants in closed-loop. While there have been a handful of works proposed towards the verification of closed-loop systems with feed-forward neural network controllers, this area still lacks attention and a unified set of benchmark examples on which verification techniques can be evaluated and compared. Thus, to this end, we present a range of closed-loop control systems ranging from two to six state variables, and a range of controllers with sizes in the range of eleven neurons to a few hundred neurons in more complex systems.

Control Systems in Manufacturing

1994 ◽  
Author(s):  
Mruthyunjaya S. Telagi ◽  
Athamaram H. Soni
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Fuzzy Logic ◽  
Control Systems ◽  
Research Trends ◽  
Future Research ◽  
Cerebellar Model Articulation Controller ◽  
Manufacturing Environment ◽  
The Future ◽  
Planar Robot

Abstract This paper reviews different control methodologies applied in manufacturing environment. Since comparatively newer control methodologies like — Neural networks, Fuzzy logic, and Cerebellar model articulation controller have gained more research interests in recent years, they have been dealt in more detail. With this, we have presented application of neural network for endeffector positioning of three degree planar robot and results have been evaluated. Finally the future research trends in these areas have been discussed.

scholarly journals A NEAT Based Two Stage Neural Network Approach to Generate a Control Algorithm for a Pultrusion System

AI ◽  
2021 ◽  
Vol 2 (3) ◽  
pp. 355-365
Author(s):  
Christian Pommer ◽  
Michael Sinapius ◽  
Marco Brysch ◽  
Naser Al Al Natsheh
Keyword(s):  
Neural Network ◽  
Control Systems ◽  
High Speed ◽  
Control Algorithm ◽  
Time Frame ◽  
Physical Processes ◽  
Neural Network Approach ◽  
Alternative Approach ◽  
Reasonable Time Frame ◽  
Speed Training

Controlling complex systems by traditional control systems can sometimes lead to sub-optimal results since mathematical models do often not completely describe physical processes. An alternative approach is the use of a neural network based control algorithm. Neural Networks can approximate any function and as such are able to control even the most complex system. One challenge of this approach is the necessity of a high speed training loop to facilitate enough training rounds in a reasonable time frame to generate a viable control network. This paper overcomes this problem by employing a second neural network to approximate the output of a relatively slow 3D-FE-Pultrusion-Model. This approximation is by orders of magnitude faster than the original model with only minor deviations from the original models behaviour. This new model is then employed in a training loop to successfully train a NEAT based genetic control algorithm.

scholarly journals Modern Problems of Information Security in Control and Access Control Systems When Using Neural Networks

2021 ◽  
Vol 2096 (1) ◽  
pp. 012159
Author(s):  
V A Chelukhin ◽  
S E Tikhonov ◽  
Pyae Zone Aung
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Information Security ◽  
Access Control ◽  
Control Systems ◽  
Theoretical Study ◽  
Management Systems ◽  
Control And Management ◽  
Network Technologies ◽  
Neural Network Technologies

Abstract This work is devoted to a theoretical study of the investigation of incidents from the operation of access control systems using neural networks in our time. The work describes the processes of operation of control and access control systems, in which neural network technologies are most actively introduced among the components of access control and management systems, and which, from the introduction of neural networks into them, can show new vulnerabilities in the operation of the access control and management system as a whole.

Methods and technologies of machine learning in neural network for computer vision purposes

2021 ◽  
Author(s):  
A.A. Adamova ◽  
V.A. Zaykin ◽  
D.V. Gordeev
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Neural Networks ◽  
Computer Vision ◽  
Human Life ◽  
Learning Technologies ◽  
Convolutional Networks ◽  
Emergency Situations ◽  
Control Functions ◽  
The Moment

This article is devoted to an overview of the current state and development prospects in the field of machine learning technologies application in computer vision problems. The article discusses the types of architectures of deep convolutional networks used for image processing, discusses their application in the space industry and provides an analysis of the element base for the implementation of computer vision platforms. The aim was to research the machine learning methods in computer vision problems. Consideration of options for using neural networks in solving problems related to astronautics. The authors considered various methods and technologies of machine learning using both domestic and foreign devices. The study showed that at the moment there are several domestic companies that are engaged in the development of microprocessors, on which it is possible to implement a neural network and train it. Also, the prospects of machine learning in computer vision problems, their possibility and feasibility of application at the present time and in the near future were identified. The results of the work can be used to create various types of neural networks. Based on the above overview of neural processors, you can begin to design a neural network. The processing and dumping of incoming information, necessary for machine learning, is able to control functions, solve emergency situations and protect human life.

Neural Networks and Identification of Systems With Unobserved States

1993 ◽  
Vol 115 (1) ◽  
pp. 196-203 ◽  
Author(s):  
C. J. Goh ◽  
Lyle Noakes
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Control System ◽  
Nonlinear Control ◽  
Control Systems ◽  
Output Feedback ◽  
Original System ◽  
Nonlinear Control Systems ◽  
Output Feedback Controller ◽  
Network Emulator

Consider a nonlinear control system, whose structure is not known (apart from the order of the system) and whose states are not observed. We observe the output of the system for a period of time using persistently exciting input, and use the observation to train a neural network emulator whose output approximates that of the original system. We point out that such an explicit dynamical relationship between the input and the output is useful for the purpose of construction of output feedback controller for nonlinear control systems. Specialization of the method to linear systems allows swift convergence and parameter identification in some cases.

scholarly journals Design of Satellite Attitude Control Systems using Adaptive Neural Networks

2020 ◽  
Vol 12 (3) ◽  
pp. 173-182
Author(s):  
M. RAJA ◽  
Kartikay SINGH ◽  
Aishwerya SINGH ◽  
Ayush GUPTA
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Adaptive Control ◽  
Control Systems ◽  
Attitude Control ◽  
Research Work ◽  
Poor Performance ◽  
Error Feedback ◽  
Pid Tuning ◽  
Adaptive Neural Networks

This paper investigates the performance of adaptive neural networks through simulations for satellite systems involving three-axis attitude control algorithms. PID tuning is the method employed traditionally. An optimally tuned, to minimizes the deviation from set point. It also responds quickly to the disturbances with some minimal overshoot. However, the disadvantage of poor performance has been observed in these controllers when manual tuning is used which in itself a monotonous process is. The PID controller using Ziegler-Nichols has more transient responses of satellite such as Overshoot, Settling time, and Steady state errors. For overcome this technique, the proposed analysis implemented an Adaptive Neural Network with PID tuning. The paper aims to combine two feedback methods by using neural networks. These methods are feed- forward and error feedback adaptive control. The research work is expected to reveal the inside working of these neural network controllers for state and error feedback input states. An error driven adaptive control systems is produced, when the neural networks acquire the knowledge of slopes and gains regarding the error feedback, while, with state feedback the system will keep trying to approximate a stable approach in order to stabilize the attitude of the satellite.

Sign in / Sign up

Export Citation Format

Share Document