NEUROFUZZY SYSTEM DESIGN FOR CONTROL APPLICATIONS
Part1: Theory and Design
Abstract:
Rule-based fuzzy control, in which the plant model is replaced by a number of control rules, given by an expert or deduced by observation, provides an alternative approach for building a model and has been developed significantly in the last few years. On the other hand, the potential benefits of neural networks extend beyond the high computation rates provided by the massive parallelism to provide a greater degree of robustness, fault tolerance, adaptivity and generalization. Seeking for integrating these two approaches brings what is so-called neurofuzzy system which gives rise to gain the merits of both approaches while avoiding their individual drawbacks.
Structural and functional mapping from a fuzzy logic-based algorithm to the neural network-based approach has been considered with a thorough design procedures for both SISO and MIMO control systems. In the former, an iterative input-output gains setting structural neurofuzzy controller has been established based on a fuzzifier element of triangular equation form, a fixed fuzzy production rules learned by back-propagation neural network and a centre of gravity defuzzificztion element, while in the latter the input-output data collected from the structural form has been described by their centres only, which has been extracted based on fuzzy number using the mean-tracking clustering algorithm, to be learned by a backpropagation neural network while leaving their widths to be handled by the interpolation feature offered by this network.
NEUROFUZZY SYSTEM DESIGN FOR CONTROL APPLICATIONS
Part2: Applications
by: Mohammed M. Ali, Dr. Yousif M. Al-Assaf and Dr. Kais B. Mirza
Abstract:
While the first part of this paper was concerned primarily with the issues of neurofuzzy controllers [ structural and functional ] theory and design, the second part presents the nonlinear SISO inverted pendulum and the MIMO industrial compressor control problems as a control applications for these neurofuzzy controllers. By defining a performance measure, the behaviour of the proposed controllers in terms of the learning ability, robustness and generalization are evaluated through a number of simulation studies along with other controllers as a comparative study. Some important conclusions are drawn from these studies.
Authors: Kasim Al-Aubidy, Mohammed Ali
Title: A Hierarchical Neuro-Fuzzy MRAC of a Robot in Flexible Manufacturing Environment
Abstract: In one hand, the Model Reference Adaptive Control (MRAC) architecture has been widely used in linear adaptive control field. The control objective is to adjust the control signal in a stable manner so that the plant’s output asymptotically tracks the reference model’s output. The performance will depend on the choice of a suitable reference model and the derivation of an appropriate learning scheme. While in the other hand, clusters analysis has been employed for many years in the field of pattern recognition and image processing. To be used in control the aim is being to find natural groupings among a set of collected data. The mean-tracking clustering algorithm is going to be used in order to extract the input-output pattern of rules from applying the suggested control scheme. These rules will be learnt later using the widely used Multi-layer perceptron neural network to gain all the benefits offered by those nets. A hierarchical neuro-fuzzy MRAC is suggested to control robots in a flexible manufacturing system. This proposed controller will be judged for different simulated cases of study to demonstrate its capability in dealing with such a system.
Keywords: MRAC, mean-tracking clustering algorithm, MLP neural nets, computer control, real-time systems, robots, FMS
July 2004, The International Arab Journal of Information Technology, Volume and Issue: Vol. 1, No. 2, Website: https://www.iajit.org publisher: Zarqa University, Jordan, pp: 209-214
Authors: Mohammed Mahdi Ali
Title: Design with Simulation of an Iterative Fuzzy Logic Controller (IFLC)
Abstract: As the need for control is extended to systems of increasing complexity which is also often highly non-linear, rule-based fuzzy logic control was one of the successful solutions. An Iterative Fuzzy Logic Controller (IFLC) is designed based on finding the best input-output scale factors settings that are found during iterations to give an acceptable controlled behavior. To verify the ability of the proposed IFLC, a promising simulation study on different kinds of control systems has been pointed out
Keywords: Iterative Fuzzy Logic Controller (IFLC), Fuzzy Production Rules (FPRs), Input-Output scale factors (Ge, Gce, and Gu),
2006, Asian Journal of Information Technology, Volume and Issue: 5, 1, Year: 2006, Website: www.medwelljournals.com, publisher: Medwell Online
pp: 34-37
Authors: Omar R. Daoud, Mohammed M. Ali
Title: Wireless Cellular Systems Performance Improvement Based on Neural Network
Abstract: In this paper, a neural network (NN) part has been imposed to overcome a previously mitigated drawback that is found in Orthogonal Frequency Division Multiplex technology (OFDM) systems. In the learning process we make use of the results obtained from the previously published work to reduce the Peak to Average Power Ratio (PAPR) problem based on different linear coding techniques. The proposed technique shows that an improvement in the OFDM technology performance has been achieved based on reducing the system’s complexity. Moreover, the reduction percentage of the PAPR compared to the previously published one; which combats the PAPR based on Low Density Parity Check (LDPC), turbo coding and convolutional coding has been attained exactly. Our simulation results show that 15% reduction in PAPR over current values in the literature can be achieved depending on the system’s type. This is in addition to that the use of NN reduces the overall OFDM system's complexity. This is because that in the proposed technique the system does not need to send extra data to recombine the processed OFDM symbols at the receiver side. Thus, the performance improvement could be attained. Keywords—Multiple Input Multiple Output, Orthogonal Frequency Division Multiplexing, Neural Network, Linear codes.
Keywords: Multiple Input Multiple Output, Orthogonal Frequency Division Multiplexing, Neural Network, Linear codes.
2012, INTERNATIONAL JOURNAL OF COMMUNICATIONS, Volume and Issue: 6, 4, Website: https://ijoc.org/index.php/ijoc, publisher: USC University of Southern California
pp: 145-152
Authors: Q. J. Hamarsheh, O. R. Daoud, M. M. Ali, A. A. Damati
Title: Efficiency enhancement based on allocating bizarre peaks
Abstract: A new work has been proposed in this paper in order to overcome one of the main drawbacks that found in the Orthogonal Frequency Division Multiplex (OFDM) systems, namely Peak to Average Power Ratio (PAPR). Furthermore, this work will be compared with a previously published work that uses the neural network (NN) as a solution to remedy this deficiency. The proposed work could be considered as a special averaging technique (SAT), which consists of wavelet transformation in its first stage, a globally statistical adaptive detecting algorithm as a second stage; and in the third stage it replaces the affected peaks by making use of moving average filter process. In the NN work, the learning process makes use of a previously published work that is based on three linear coding techniques. In order to check the proposed work validity, a MATLAB simulation has been run and has two main variables to compare with; namely BER and CCDF curves. This is true under the same bandwidth occupancy and channel characteristics. Two types of tested data have been used; randomly generated data and a practical data that have been extracted from a funded project entitled by ECEM. From the achieved simulation results, the work that is based on SAT shows promising results in reducing the PAPR effect reached up to 80% over the work in the literature and our previously published work. This means that this work gives an extra reduction up to 15% of our previously published work. However, this achievement will be under the cost of complexity. This penalty could be optimized by imposing the NN to the SAT work in order to enhance the wireless systems performance.
Keywords: Orthogonal Frequency Division Multiplexing, Neural Network, Linear Codes, Wavelet, Moving Average Filter.
August 2016, International Journal of Wireless Mobile Networks (IJWMN), Volume and Issue: 8, 4, Website: http://airccse.org/journal/ijwmn.html, publisher: AIRCC Publishing Corporation, pp: 107-118
Authors: Q. J. Hamarsheh, O. R. Daoud, M. M. Ali, A. A. Damati
Title: Power Peaks Allocation Based on Averaging-Adaptive Wavelet Transform
Abstract: One of Orthogonal Frequency Division Multiplexing deficiency has been taken into consideration in
this work. A proposition has been made to tackle the Peak to Average Power Ratio (PAPR) problem. The proposed work will be based on a special averaging adaptive wavelet transformation (SAAWT) process. It will be compared with two main works that has been published previously; a neural network (NN)-based and a special averaging technique (SAT)-based.
In the NN work, the learning process makes use of a previously published work that is based on three linear coding techniques. The proposed work (SAAWT) consists of three main stages; extracting the needed features, de-noising and the optimization criterion. SAAWT has an enhancement over the SAT that will take the noise clearance enhancement into its consideration. It uses 136880 different combinations of de-noising parameters that are experimentally computed to get the most efficient result with respect to the MSE, SNR and PSNR values. A MATLAB simulation-based of such works has been made in order to check the proposition performance. In this simulation, both of the BER and CCDF curves have been taken into consideration. Furthermore, the bandwidth and channel behaviors have been remaining constant. Moreover, two kinds of data have been imposing to this
simulation; a random data that is generated randomly by making use of the MATLAB features and a practical data that have been
extracted from a funded project entitled by ECEM. From the previously published work the SAT shows promising
results in reducing the PAPR effect reached up to 75% over the work in the literature and over the NN-based work. Under the cost of increasing complexity, SAAWT gives further reduction over the SAT reaches up to 6%. This drawback will be examined in the future work.
Keywords: Orthogonal Frequency Division Multiplexing, Neural Network, Linear Codes, De-noising Parameters, Wavelet,
Moving Average Filter,
2016, International Journal of Circuits, Systems and Signal Processing, Volume and Issue: 10, Website: https://www.naun.org/cms.action?id=3029, publisher: ISSN
pp: 440-447
Authors: Mokhles M. Abdulghani Kasim M. Al-Aubidy Mohammed M. Ali and Qadri J. Hamarsheh
Title: Wheelchair Neuro Fuzzy Control and Tracking System Based on Voice Recognition
Abstract: Autonomous wheelchairs are important tools to enhance the mobility of people with disabilities. Advances in computer and wireless communication technologies have contributed to the provision of smart wheelchairs to suit the needs of the disabled person. This research paper presents the design and implementation of a voice controlled electric wheelchair. This design is based on voice recognition algorithms to classify the required commands to drive the wheelchair. An adaptive neuro-fuzzy controller has been used to generate the required real-time control signals for actuating motors of the wheelchair. This controller depends on real data received from obstacle avoidance sensors and a voice recognition classifier. The wheelchair is considered as a node in a wireless sensor network in order to track the position of the wheelchair and for supervisory control. The simulated and running experiments demonstrate that, by combining the concepts of soft-computing and mechatronics, the implemented wheelchair has become more sophisticated and gives people more mobility.
Keywords: wheelchair control; voice recognition; autonomous wheelchair; ANFIS; V-REP; mechatronics
19- 5 -2020, Sensors, Volume and Issue: 20, 10, Year: 2020, Website: https://doi.org/10.3390/s20102872, publisher: MDPI.
Authors: F. Mohammed, Wagah; M. Ali, Mohammed; N. Al-Tikriti, Munther; Kaleel, Kalid
Title: THE EFFECTS OF MOS LAYERS ON SENSING PROPERTIES OF MOS PHOTOSENSOR
Abstract: In this research work, many samples of metal oxide silicon photo sensors were laboratory prepared by thermal evaporation techniques. Some silicon samples were left in the air for a predefined time for SiO2 to grow naturally, while others were thermally coated with measured thickness of SiO. A number of the samples were coated with nickel while others with aluminum and one sample was coated with indium. Various tests and measurements were conducted; these include transmittance tests with a range of wavelength and for different thicknesses. The ideality factors of the samples and the potential barrier height were calculated from I-V and C-V characteristics. The photo generated current of the samples were also measured at photoconductive mode under reverse voltage. Quantum efficiency measurement indicated that native oxide samples provided higher quantum efficiency than those thermal
Key words: Schottky Barrier Diode, Photo sensor, MOS Photo sensor, Silicon Photo sensor:
5/6/2013, INTERNATIONAL JOURNAL ON SMART SENSING AND INTELLIGENT SYSTEMS, Volume and Issue: 6, 3, Website: https://www.researchgate.net/publication/287569295_The_effects_of_mos_layers_on_sensing_properties_of_mos_photosensor, publisher: ISSN, pp: 1102-1110
Authors: W. F. Mohammad; K. M. Al-Aubidy; M. M. Ali
Title: Maximum power point neuro-fuzzy tracker for photovoltaic arrays
Abstract: Every photovoltaic (PV) array has unique point at which maximum power can be generated and extracted for different atmospheric conditions and output load. The maximum power point tracking is therefore critical for the success of PV arrays. All methods proposed earlier suffer from slow response; they oscillate around the maximum power point and ignoring the atmospheric conditions changing. This paper presents an intelligent method of maximum power point tracking for photovoltaic systems. It is based on tracking the maximum power point by monitoring the voltage and current of the solar array and adjusting the duty cycle of the PWM switching signal of a buck-boost DC/DC converter. Both conventional fuzzy logic controller and neuro-fuzzy controller are implemented to evaluate PV system performance. Functional neuro-fuzzy controller has advantages over that of fuzzy logic both in speed and generalization features. Obtained results show that the neuro-fuzzy controller can deal with different load and weather conditions and deliver more power from the photovoltaic systems.
Key words: Photovoltaic System, Solar cells, Fuzzy logic, Neuro-fuzzy, Maximum power point tracking
2011, Eighth International Multi-Conference on Systems, Signals Devices, Tunisia, Date of conference: 22-25 March 2011, Vo.43, Issue24, Pages: 30-35
Publisher: IEEE
Authors: Kasim M. Al-Aubidy; Mohammed M. Ali; Ahmad M. Derbas; Abdallah W. Al-Mutairi
Title: GPRS-based remote sensing and teleoperation of a mobile robot
Abstract: The main objective of this research was to design and implement a remote sensing and monitoring system running on mobile robot with obstacle avoidance capability in unreachable area. A simple mobile robot prototype with onboard sensors has been designed and implemented to scan and monitor several variables in the surrounding environment. Teleoperation of such a mobile robot i s a challenging task that requires an efficient interface and a reliable real-time robot control to avoid obstacles. The proposed system enables the user (base station) to send commands to the remote station (mobile robot), and receive scanned data and images from the environment through the internet and mobile DTMF signal. The proposed system hardware and software was implemented using PROTUS development software to obtain the suitable design parameters. Then, real experiments have been achieved to demonstrate the system performance including both the ultrasonic teleoperation of mobile robot navigation to avoid obstacles, and real-time sensing and monitoring in unreachable area.
Key words: Mobile Robot, Robot navigation, Remote sensing and monitoring, Wireless sensor networks, Obstacles avoidance.,
2018, Sensors, Circuits Instrumentation Systems, Tunisia, Date of conference: 19-22/March/2018, Vol: 6, Pages: 113-128, Publisher: IEEE
Authors: Kasim M. Al-Aubidy; Mohammad M. Ali; Ahmad M. Derbas; Abdallah W. Al-Mutairi
Title: Real-time monitoring and intelligent control for greenhouses based on wireless sensor network
Abstract: The main objective of this research is to design and implement a real-time monitoring and control of several environmental parameters for group of greenhouses. Each greenhouse is considered as a node in a wireless sensor network. A single-board microcontroller-based system has been designed and implemented to monitor and control several variables and maintain desired condition in each greenhouse. A rule-based fuzzy controller has been designed to control the microclimate of each greenhouse. The proposed system enables the farmer to monitor both the internal environment of the greenhouse. Also, the farmer can send commands to turn ON or OFF certain devices in a selected greenhouse through wireless communications. Simulated and real results have been achieved to demonstrate the system performance and real-time remote monitoring and control activities.
Key words: Greenhouse automation, Remote monitoring and control, Fuzzy control, Intelligent control, Wireless sensor network.,
2014, 11th International Multi-Conference on Systems, Signals Devices (SSD14), Barcelona, Spain, 11-14/2/2014, Vol.1, Issue:1, Pages:1-7, Publisher: IEEE
Authors: Ahmad M. Derbas, K. Al-Aubidy, M. M. Ali, Abdallah W. Al-Mutairi
Title: Multi-robot system for real-time sensing and monitoring
Abstract: The main objective of this research is to design and realize a multi-robot system for real-time sensing and monitoring suitable for hazardous and/or unreachable environment. The proposed system has three mobile robots; main, rover and eye. Each mobile robot has its own embedded microcontroller and set of sensors. Wireless communications between local site and these mobile robots are achieved by WiFi, ZigBee and Bluetooth techniques, and can be accessed through the internet. Wireless teleoperation of these mobile robots is a challenging task that requires an efficient interface and a reliable real-time control algorithm to avoid obstacles. The proposed system enables the authorized operator to send commands to the mobile robots, and receive scanned data and images from the environment through the internet. The mechanical part of the remote station has been built after careful selection of the design parameters using CAD/CAM tools. While, the system hardware and software parts of the embedded controllers were implemented using PROTEUS development tool to obtain the suitable design parameters. Then, real experiments have been achieved to demonstrate the system performance including the wireless teleoperation of the three mobile robots, their navigation to avoid obstacles, and real-time sensing and monitoring.
Key words: Mobile robot, Multi-robot system, Remote sensing and monitoring, Wireless sensor networks, Obstacles avoidance.,
2014, 15th International Workshop on Research and Education in Mechatronics (REM), Egypt, Publisher: IEEE
Authors: Tariq Younis Ali; Mohammad M. Ali
Title: Robotino obstacles avoidance capability using infrared sensors
Abstract: Most of the recent mobile robot researchers focus on obstacle avoidance and path tracking in unknown environment. This paper presents a new algorithm using Straight-Line Equation adaptation mechanism that makes Robotino reaching its destination accurately, also to enable it to detect and to avoid static or dynamic obstacles using nine infrared sensors. A brief Robotino dynamic description is discussed to help in understanding the proposed control algorithm. A detailed algorithm design procedure is evaluated. The simulation results showed the effectiveness of the proposed algorithm in the sense of avoiding obstacles without collision through Robotino predefined path.
Key words: Robotino, Obstacles avoidance, infrared sensors, Year:2015, Name of conference: 2015 IEEE Jordan Conference on Applied Electrical Engineering and Computing Technologies (AEECT), Place of conference: Jordan, Date of conference: 3-5 Nov. 2015, Publisher: IEEE
Authors: Mohammad M. Ali; Ali Al-Khawaldeh
Title: A simulation study of multi-disciplinary position control methods of robot arm D.C motor
Abstract: This paper presents a simulation study using different control strategies to control the position of robot arm DC motor. Fixed field DC motor mathematical model is applied using certain parameters settings. A state feedback pole placement, Fuzzy Logic, Multi-Layer Perceptron (MLP) Neural Network, and the conventional PID control theories have been applied successfully. Matlab Simulink work space is used in the simulation. Almost the same controlled output responses are obtained with a different transient responses speed.
Key words: PID, MLP NN, multi-disciplinary position control,
2016, 13th International Multi-Conference on Systems, Signals Devices (SSD), Place of conference: Germany, 21-24 March 2016., Pages: 489-493, Publisher: IEEE
Authors: Mohammed Rabeea Hashim Al-Dahhan ; Mohammad M. Ali
Title: Path tracking control of a mobile robot using fuzzy logic
Abstract: Recently, the study and development of the mobile robot is considered as a very important issue for many researchers. This is because the wide range of mobile robot applications in real life. One of the most important mobile robot tasks is the control of its navigation in tracking its predefined path. This also need a good capability in avoiding any static or dynamic obstacles that the mobile robot faces in its route until reaching its destination. The difficulty in finding a good mathematical model for the mobile robot used in this research "Robotino® from Festo company" made the decision to use fuzzy logic to design a controller capable to introduce a safe Robotino® navigation. Fuzzy logic controller needs information about Robotino® features and behavior in order to build its rule base which are inspired from human experience in such application. These rules can be easily programmed to bring out an efficient controller. Sugeno algorithm is implemented which the experiments results validated its efficiency. Fuzzy logic controller with 153-fuzzy rule is used for controlling the Robotino® path tracking issue, while another fuzzy logic controller with 27-fuzzy rule is applied for the Robotino® obstacle avoidance feature. Matlab is used as a tool to implement the two proposed fuzzy controllers. Many real-time experiments have been conducted in the Faculty of Engineering research laboratory at Philadelphia University. Results reflect the good abilities of the proposed controllers.
Key words: Mobile Robot, Robotino®, Path Tracking, Obstacle Avoidance, Fuzzy Logic Controller.
2016, 13th International Multi-Conference on Systems, Signals Devices (SSD), Germany, 21-24 March 2016, Pages: 83-88, Publisher: IEEE
Authors: Mohammad M. Ali; Tariq Younis Ali
Title: Obstacles avoidance for omnidirectional mobile robot using line trajectory adaptation
Abstract: This paper presents an algorithm which is designed based on the adaptation of Straight-Line Equation parameters in order to detect and avoid both static and dynamic obstacles. A real-time measurement is collected making use of the already built-in nine infrared sensors along with the added ultrasonic sensor to increase the obstacle recognition range. The related control actions coming from the executing of the control algorithm are used to force the mobile robot movement through its three drive units to reach destination safely. This has been achieved by updating the required distance and orientation angle. The experimental results showed the effectiveness of the proposed algorithm in the sense of avoiding obstacles without collision and reaching the goal with minimum position error.
Key words: Mobile Robot, Robotino®, Obstacle Avoidance, line trajectory adaptation, 2016, 13th International Multi-Conference on Systems, Signals Devices (SSD), Germany, 21-24 March 2016, Pages: 96-101, Publisher: IEEE
Authors: Mustafa Hatem Jebur ; Mohammad M. Ali
Title: Safe navigation and target recognition for a mobile robot using neural networks
Abstract: Recently, there is a large demand on using mobile robots in different life applications. Thus, it is of importance to ensure mobile robot safe navigation towards its destination. In this research Robotino® from Festo company is used to confirm safe navigation issue along with red color target recognition using its IR and camera sensors respectively. Real-time and simulation experimental results have been obtained in laboratories of faculty of engineering / Philadelphia university / Jordan. Results were taken to train two multi-layers’ perceptron neural networks. One of them is used to force Robotino® moving towards its target by controlling its linear velocity, while the other one is used to move Robotino® avoiding any possible static or dynamic obstacle in its route. Matlab workspace is used for system analysis and design.
Key words: Safe navigation, Robotino, Neural Networks,
2017, 2017 14th International Multi-Conference on Systems, Signals Devices (SSD), Morocco
28-31 March 2017, Pages: 705-712, Publisher: IEEE