Quadcopter Stabilization using Neural Network Model from Collected Data of PID Controller

In the present study, deep learning neural network model has been employed in many engineering problems including heat transfer prediction. The main consideration of this document is to predict the performance of the boiling heat transfer in helical coils under terrestrial gravity conditions and compare with actual experimental data. Total of 877 data sample has been used in the present neural model. Artificial new Neural Network (ANN) model developed in Python environment with Multi-layer Perceptron (MLP) using four parameters (helical coils dimensions, mass flow rate, heating power, inlet temperature) and one parameter (outlet temperature) has been used in the input layer and output layer in order. Levenberg-Marquardt (LM) algorithm using L2 Regularization to find out the optimal model. A typical feed-forward neural network model composed of three layers, with 30 numbers of neurons in each hidden layer, has been found as optimal based on statistical error analysis. The 4-30-30-1 neural model predicts the characteristics of the helical coil with the accuracy of 98.16 percent in the training stage and 96.68 percent in the testing stage. The result indicated that the proposed ANN model successfully predicts the heat transfer performance in helical coils and can be applied for others operation concerned with heat transfer prediction for future works

In this paper, a reconfiguration approach using fuzzy logic algorithm and neural network modelling is proposed. When failure has been detected, the state of the degraded system is evaluated by comparing the output of the system with the estimation provided by a neural model. Therefore, we propose to use the neural network in order to cover all the operating zone of the faulty system. By combining neural network capabilities and fuzzy logic for fault evaluation, a new control law is determined taking into account the impact of the failure on the system. Its potentialities are illustrated through simulation studies on a thermal plant presenting bilinear characteristics. >

Single-input interval type-2 fuzzy PID (SI-IT2-FPID) control approach is a simple and effective way of adding nonlinear characteristics to linear PID controllers to increase the control performance and the capability of handling system nonlinearities. In the SI-IT2-FPID approach, the nonlinear characteristic of the PID controller is provided by a single-input–single-output interval type-2 fuzzy controller (IT2-FC) positioned before the PID controller. In the literature, a constant footprint of uncertainty (C-FOU) structure is used in SI-IT2-FPID controller design procedures in almost all studies. However, the C-FOU structure limits the performance of the SI-IT2-FPID controller. In this study, an online footprint of uncertainty (FOU) tuning mechanism is proposed for SI-IT2-FPID controllers in order to increase their control performances. Online tuning of FOUs of interval type-2 membership functions (IT2-MFs) enhances the capability of SI-IT2-FPID controllers to handle system nonlinearities since FOU levels of IT2-MFs have a dominant effect on the control and robustness performance of SI-IT2-FPID controllers. In the proposed method, FOUs of IT2-MFs of an IT2-FC are adjusted by an analytical function based on the system error and the change of the error. In this way, the simple structure of the SI-IT2-FPID controller is preserved. In order to illustrate the control and robustness performance of the proposed controller, simulation studies are carried out on three benchmark systems, namely a second-order nonlinear system, a ball and beam system, and a first-order plus dead time system. Additionally, the performance of the proposed controller is compared to a SI-IT2-FPID controller with C-FOU and a classical PID controller. The simulation results show that the proposed controller exhibits better control and robustness performance than the other controllers.

In the field of autonomous robotics, enhancing the navigation system of robots is a crucial aspect that directly impacts their performance. This study presents a novel approach to addressing this challenge with an Artificial Neural Network (ANN) model. The research focuses on improving the navigation capabilities of a differential drive robot using the line-following method for route tracking and the dead reckoning technique for localization. It investigates a differential drive robot model controlled with a PID controller and derives the transfer function of the PID model. Through simulations, it becomes apparent that the PID model exhibits a continuous overshoot in its response, which negatively affects the behaviour of the robot's wheels. Ordinarily, continuous manual tuning will be required to correctly tune the PID controller to a value where the overshoots will be negligible, and this could be onerous. To overcome this limitation, an ANN controller is proposed, leveraging the learning capabilities of the neural network. Data from the PID controller transfer function is utilized to train the ANN model, enabling it to understand patterns and relationships. The ANN controller is then substituted in place of the PID controller in the simulation. The results showcase a remarkable 13.1% improvement in the robot's wheel response, highlighting the transformative potential of this approach for revolutionizing autonomous robot navigation in industrial applications. By using the transfer function of the PID model to train an ANN model, this study offers a powerful framework for enhancing the navigation performance of a differential drive autonomous robot and shows performance improvements in control, flexibility, and adaptation to changing conditions. These discoveries have significant ramifications for the industry and will pave the way for intelligent and effective autonomous robot navigation systems. The research provides a comprehensive understanding of the challenges associated with the differential drive robot model controlled with a PID controller and offers a robust approach to how this can be alleviated. The significance of this study lies in its ability to address the continuous overshoot issue observed in the PID controller's response by training an ANN controller with data from the PID controller. The proposed approach minimizes overshoot and improves the robot's wheel response, ultimately enhancing its navigation capabilities. Overall, this study demonstrates the potential of an ANN model to revolutionize autonomous robot navigation in industrial applications. The notable improvement achieved in the robot's wheel response validates the effectiveness of this approach. Future research can further optimize this integrated approach in real-world scenarios, leading to intelligent and efficient autonomous robot navigation systems across diverse industrial settings.

Under the research context of two axis digital control turntable, this paper has set up a mathematical model of control system of turntables. Proceeding from such model, it has made an analysis on the deficiencies of classical PID control during turntable control. Against the feature of PID parameters difficult to be tuned, the neural network is thus introduced in the optimizing process of turncontrol PID parameters, during which PID controller is designed based on BP network. In addition, the PID algorithm control based on BP network is simulated with results indicating PID controller designed so featuring high tracking accuracy, stable performance and strong robustness, which provides a solution to the problem of inadequate precision required by classic PID control as regards to the control precision.

Event Abstract Back to Event A layered description language for spiking neural network modeling Chung-Chuan Lo1* 1 National Tsing Hua University , Institute of Bioinformatics and Structural Biology, Taiwan The neural modeling has a long history in the development of modern neuroscience and has greatly enhanced our understanding in principles of neural activity and interactions between neurons. However, as a rapidly growing field, the neural network modeling has reached a level of complexity which makes the exchange of information between research groups extremely difficult. It also becomes more and more unlikely that the modeling results obtained in one lab be exactly reproduced in another lab which uses a different simulator. The problem arises from the fact that the field of computational neuroscience is lacking appropriate standards to communicate network models. To address this issue, the International Neuroinformatics Coordinating Facility (INCF) has initiated a project: Network Interchange for Neuroscience Modeling Language, or NineML, which provides a standardized machine-readable language for spiking neural network models with an aim to ease model sharing and to facilitate the replication of results across different simulators. In the talk I will introduce the first version of NineML. Its most innovated features includes: 1. Layered: The complete description of a neural network model in NineML is separated into to a user layer and an abstraction layer. The XML-based user layer provides a syntax to specify the instantiation and parameterization of a network model in biological terms. The abstraction layer provides explicitly descriptions of the core concepts, mathematics, model variables and state update rules. 2. Fully self-consistent: All model concepts defined in the user layer are expressed explicitly in the abstraction layer so that a neural network model can be unambiguously implemented by software that fully supports NineML. 3. Highly expandable: future expansions are taken into account in the development of NineML. Hence specific model features that are not supported in the current version of NineML can be easily added in a later version without any major revision to the specification of the language. In the talk I will also demonstrate NineML using several example models of neural networks. I will show how the description looks like in different layers and how NineML solves some difficult problems. Using NineML, researchers can describe their neural network models in an unambiguous and simulator-independent way. Furthermore, the models can be reimplemented and simulation results can be easily reproduced by any simulator which fully supports NineML. We believe that this project will have a profound effect on the modeling community and will facilitate research in computational neuroscience. Conference: Neuroinformatics 2010 , Kobe, Japan, 30 Aug - 1 Sep, 2010. Presentation Type: Oral Presentation Topic: Workshop 1: How to describe a model: description language solutions and challenges Citation: Lo C (2010). A layered description language for spiking neural network modeling. Front. Neurosci. Conference Abstract: Neuroinformatics 2010 . doi: 10.3389/conf.fnins.2010.13.00001 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 08 Jun 2010; Published Online: 08 Jun 2010. * Correspondence: Chung-Chuan Lo, National Tsing Hua University, Institute of Bioinformatics and Structural Biology, Hsinchu City, Taiwan, cclo@life.nthu.edu.tw Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Chung-Chuan Lo Google Chung-Chuan Lo Google Scholar Chung-Chuan Lo PubMed Chung-Chuan Lo Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

The influence of the inlet and outlet heights of the flow channel on the performance of a proton exchange membrane fuel cell was studied with a three-dimensional fuel cell mathematical model. The numerical results were used as the basic training data of a three-layer back-propagation artificial neural network (ANN) model in which the inlet and outlet heights of the flow channel as well as the cell voltage were set to be the input variables while cell current density is the output variable. Then, applying the Bagging ensemble learning method to integrate the neural network model, a fuel cell performance prediction method is constructed. It was shown that as compared with the ANN model, the Bagging neural network model exhibited higher prediction accuracy and less requirements of model training data. It also can be applied for rapid prediction of fuel cell performance in a wide range of conditions beyond the training data, showing the robustness of the proposed Bagging neural network ensemble model.

In this paper, the Direction of Arrival -DoA estimation for two mobile sources was performed by using the Single Multilayer Perceptron (MLP) neural network model (SMLP-DoA) and the Cascade MLP model(CMLP).The latter model consists of two neural networks connected in a cascade where the outputs of the first MLP that rejects noise represent the inputs to the second network in a cascade.The outputs of the neural network models determine the direction of arrival of the incoming signals.Two cases were considered, in the first case the neural networks were trained on the samples that were without noise, and in the second with samples containing noise.Both considered neural network models were tested with noisy samples.The results of these two neural models are compared to the results achieved by the RootMUSIC algorithm.The presented results show that the proposed CMLP model has a higher accuracy in determining the angular positions of sources compared to the classical SMLP-DoA model and the RootMUSIC algorithm.Moreover, the CMLP model executes significantly faster compared to the model based on the RootMUSIC algorithm.

In this paper, a speech-to-text translation model has been developed for Malaysian speakers based on 41 classes of Phonemes. A simple data acquisition algorithm has been used to develop a MATLAB graphical user interface (GUI) for recording the isolated word speech signals from 35 non-native Malaysian speakers. The collected database consists of 86 words with 41 classes of phoneme based on Affricatives, Diphthongs, Fricatives, Liquid, Nasals, Semivowels and Glides, Stop and Vowels. The speech samples are preprocessed to eliminate the undesirable artifacts and the fuzzy voice classifier has been employed to classify the samples into voiced sequence and unvoiced sequence. The voiced sequences are divided into frame segments and for each frame, the Linear Predictive co-efficients features are obtained from the voiced sequence. Then the feature sets are formed by deriving the LPC features from all the extracted voiced sequences, and used for classification. The isolated words chosen based on the phonemes are associated with the extracted features to establish classification system input-output mapping. The data are then normalized and randomized to rearrange the values into definite range. The Multilayer Neural Network (MLNN) model has been developed with four combinations of input and hidden activation functions. The neural network models are trained with 60%, 70% and 80% of the total data samples. The neural network architecture was aimed at creating a robust model with 60%, 70%, and 80% of the feature set with 25 trials. The trained network model is validated by simulating the network with the remaining 40%, 30%, and 20% of the set. The reliability of trained network models were compared by measuring true-positive, false-negative, and network classification accuracy. The LPC features show better discrimination and the MLNN neural network models trained using the LPC spectral band features gives better recognition.

Fractional order PID, which involve integration and differentiation of non-integer order, is increasingly being used in the fields of control system, robotics, signal processing and circuit theory. Based on neural network, this paper introduce a new approach for design fractional order PID controller. A self-learning PID controller with five dimension parameters is realized by using parameter turning strategy of RBF neural network. Further apply the fractional order PID controller to a typical FOLPD process like pulp consistency control system. In particular, pulp consistency system is a typical FOLPD process. The simulation results show that this kind of controller has a better effect for pulp consistency system than traditional PID and PIDNN controller.

This paper presents an application of the neural network (NN) model for forecasting and filtering problems. Instead of using the gradient descent method as used by the standard back-propagation network, the present study adopted the Levenberg-Marquardt algorithm. It was shown that the Levenberg-Marquardt algorithm tends to reach termination criteria more quickly making the searches employing the full model less time consuming. In the first part of forecasting problem, multilayer feedforward (MLFF) network was constructed by trials to forecast the tidal-level variations at the Chao Phraya river mouth in Thailand. Unlike the well-known conventional harmonic analysis, the neural network model uses a set of previous data for learning and then forecasting directly the time series of tidal levels. It was found that lead time of 1 to 24 hourly tidal levels can be successfully predicted using only a short-time hourly learning data.In the second part of filtering problem, the MLFF network was also used to establish a stage-discharge relationship for tidal river. Comparisons were made among conventional methods, i.e. by using multivariate functions (Linear and power models) and the NN model. The NN model was found to be able to model the transfer function too much higher accuracy than multiple regression analysis. The root mean square errors (RMSE) and the mean relative error (MRE) for the NN model are about 8 cm and 6.4% whereas the same errors for multiple regression analysis are 16 cm and 12%, respectively. The efficiency index for the linear, power, and NN models is 0.4, 0.7, and 0.9, respectively. In addition, the obtained stage-discharge relationship by the NN model can indicate reasonably important behavior of the tidal influences, i.e. there are stronger influences of tide during the dry period than the flood period.

In allusion to unmanned vehicle steering control of the brushless DC motor control system, traditional PID controller parameter adjustment complex, weak ability to adapt to the environment and other issues, on the basis of the analysis of internal model control and classical PID control internal corresponding relationship, comprehensive its advantages, The design uses a brushless DC motor in the steering control system for unmanned vehicles based on the internal model PID controller ( IMC-PID) for speed. Based on the build object theoretical model, online simulation controller show that, for the design objects, based on the internal model PID controller whether the system step response or disturbance tracking control effect can reach the classic PID control requirements, also reduces the complexity and randomness of the design parameters.

Developing a new combined model of zenith wet delay by using neural network

Neural networks are a state-of-the-art approach that performs well for many tasks. The activation function (AF) is an important hyperparameter that creates an output against the coming inputs to the neural network model. AF significantly affects the training and performance of the neural network model. Therefore, selecting the most optimal AF for processing input data in neural networks is important. Determining the optimal AF is often a difficult task. To overcome this difficulty, studies on trainable AFs have been carried out in the literature in recent years. This study presents a different approach apart from fixed or trainable AF approaches. For this purpose, the activation function cyclically switchable convolutional neural network (AFCS-CNN) model structure is proposed. The AFCS-CNN model structure does not use a fixed AF value during training. It is designed in a self-regulating model structure by switching the AF during model training. The proposed model structure is based on the logic of starting training with the most optimal AF selection among many AFs and cyclically selecting the next most optimal AF depending on the performance decrease during neural network training. Any convolutional neural network (CNN) model can be easily used in the proposed model structure. In this way, a simple but effective perspective has been presented. In this study, first, ablation studies have been carried out using the Cifar-10 dataset to determine the CNN models to be used in the AFCS-CNN model structure and the specific hyperparameters of the proposed model structure. After the models and hyperparameters were determined, expansion experiments were carried out using different datasets with the proposed model structure. The results showed that the AFCS-CNN model structure achieved state-of-the-art success in many CNN models and different datasets.

Temperature prediction of lithium-ion battery based on artificial neural network model