Simple adaptive control for SISO nonlinear systems using multiple neural networks

This paper presents a method of simple adaptive control (SAC) for nonlinear systems using neural networks. The control input is given by the sum of the output of a simple adaptive controller and the output of the neural network. The neural network is used to compensate the nonlinearity of plant dynamics that is not taken into consideration in the usual simple adaptive control. The role of the neural network is to construct a linearized model by minimizing the output error caused by nonlinearities in the control system.

This paper presents a method of continuous-time simple-adaptive control (SAC) for multi-input multi-output (MIMO) nonlinear systems using multifraction neural networks. The control input is given by the sum of the output of the simple adaptive controller and the output of the multifraction neural network is used to compensate the nonlinearity of plant dynamics that is not taken into consideration in the usual SAC. The role of the multifraction neural network is to construct a linearized model by minimizing the output error caused by nonlinearities in the control systems.

Sliding mode control (SMC) has a strong capability of controlling nonlinear systems with uncertainties. However, it requires thorough knowledge of parameters and dynamics of the controlled plant, which are difficult to be obtained or may be unknown. This will cause difficulties in calculating equivalent control law of SMC. To overcome this problem, this paper proposes an adaptive SMC using simple adaptive control (SAC) developed for single-input single-output (SISO) nonlinear systems with unknown parameters and dynamics. The role of SAC is to construct an equivalent control input of adaptive SMC. To construct a corrective control input, this paper applies the method using the sign function with a modified sliding surface. Finally, the effectiveness of the proposed method is confirmed through computer simulation.

The mutual impacts between adjacent buildings caused large-scale damage during past major earthquakes sparking interest in examining the reliability and effectiveness of various control devices for alleviating the seismic responses of adjacent buildings. In this study, seismic response attenuation of multi-story adjacent buildings connected with MR dampers is examined using the Simple Adaptive Control (SAC) method. The use of the adaptive controller is advantageous as it can cope with the structural characteristics change during the severe events such as earthquakes. The simple Bouc-Wen model is considered in the analysis to model the highly nonlinear behavior of the MR damper. The Linear Quadratic Regulator (LQR) is used to design the reference model with desired trajectories. The change in the structural characteristics is reflected as a reduction in the plant's mass and stiffness as a result of the damage in the structural system. The efficacy of the simple adaptive controller is investigated in the presence of damage and in case where the system still undamaged. The results show that connecting adjacent buildings with MR dampers driven by the Simple Adaptive Control method is quite effective in mitigating the seismic responses.

This paper presents a method of simple adaptive control (SAC) using neural networks for a class of nonlinear systems with bounded-input bounded-output (BIBO) and bounded nonlinearity. The control input is given by the sum of the output of the simple adaptive controller and the output of the neural network. The neural network is used to compensate for the nonlinearity of the plant dynamics that is not taken into consideration in the usual SAC. The role of the neural network is to construct a linearized model by minimizing the output error caused by nonlinearities in the control systems. Furthermore, convergence and stability analysis of the proposed method is performed. Finally, the effectiveness of the proposed method is confirmed through computer simulation.

In this paper, the design of a neural network (NN) based adaptive model-based event-triggered control of an uncertain single input single output (SISO) nonlinear discrete time system in affine form is presented. The controller uses an adaptive estimator consisting of a single-layer NN not only to approximate the internal dynamics of an affine nonlinear discrete-time system but also to provide an estimate of the state vector during inter event interval. The NN weights of the adaptive NN estimator are tuned in a aperiodic manner at the event trigger instants unlike periodic updates in standard adaptive neural network (NN) control. A dead zone operator is used to reset the event trigger error to zero as long as the system states continue to remain in a bounded region due to NN reconstruction errors. Lyapunov method is used to derive the event trigger condition, prove uniform ultimate boundedness (UUB) of the NN weight estimation error and system states.

In this research, the positioning control of a water hydraulic servo cylinder system with the simple adaptive control (SAC) is discussed. Conventional model reference adaptive control (MRAC) also achieves good tracking performance for cylinder control, but the controller structure is much more complicated and has less robustness for disturbance in real application. Compared with MRAC, SAC has simpler and lower order structure, i.e. higher feasibility. In this paper, we examined and evaluated the control performance of SAC on a water hydraulic servo cylinder system. With the recent increased concerns over global environmental problems, the water hydraulic technique using pure tap water as a pressure medium has become a new drive source comparable to electric, oil hydraulic and pneumatic drive systems. This technique is also preferred due to its high power density, high safety against fire hazard in production plants, and easy availability. However, the main problems for precise control in a water hydraulic system are steady state errors and overshoot due to its large friction torque and considerable leakage flow. To compensate these effects, MRAC had been already applied and the better control performances were obtained. However, no result has been reported on SAC application for water hydraulics. To make clear the merits of SAC, tracking control performance and robustness were discussed via experimental results. It was confirmed that SAC gives (1) better tracking performance compared with PI control, and (2) comparable control precision with MRAC within 70 micrometers to the reference position and higher robustness for parameter change despite simple controller. These results ensure wider application of simple adaptive control in real mechanical system.

A method of face recognition based on fuzzy clustering and parallel neural networks

Emerging heterogeneous computing, computing at the edge, machine learning and AI at the edge technology drives approaches and techniques for processing and analysing onboard instrument data in near real-time. The author has used edge computing and neural networks combined with high performance heterogeneous computing platforms to accelerate AI workloads. Heterogeneous computing hardware used is readily available, low cost, delivers impressive AI performance and can run multiple neural networks in parallel. Collecting, processing and machine learning from onboard instruments data in near real-time is not a trivial problem due to data volumes, complexities of data filtering, data storage and continual learning. Little research has been done on continual machine learning which aims at a higher level of machine intelligence through providing the artificial agents with the ability to learn from a non-stationary and never-ending stream of data. The author has applied the concept of continual learning to building a system that continually learns from actual boat performance and refines predictions previously done using static VPP data. The neural networks used are initially trained using the output from traditional VPP software and continue to learn from actual data collected under real sailing conditions. The author will present the system design, AI, and edge computing techniques used and the approaches he has researched for incremental training to realise continual learning.

No AccessEngineering NoteAdaptive Suppression of Linear Structural Vibration Using Control Moment GyroscopesQuan Hu, Yinghong Jia and Shijie XuQuan HuSchool of Astronautics, Beihang University, 100191 Beijing, People's Republic of China, Yinghong JiaSchool of Astronautics, Beihang University, 100191 Beijing, People's Republic of China and Shijie XuSchool of Astronautics, Beihang University, 100191 Beijing, People's Republic of ChinaPublished Online:25 Apr 2014https://doi.org/10.2514/1.62267SectionsRead Now ToolsAdd to favoritesDownload citationTrack citations About References [1] D'Eleuterio G. M. T. and Hughes P. C., "Dynamics of Gyroelastic Continua," Journal of Applied Mechanics, Vol. 51, No. 2, June 1984, pp. 412–422. JAMCAV 0021-8936 Google Scholar[2] D'Eleuterio G. M. T. and Hughes P. C., "Dynamics of Gyroelastic Spacecraft," Journal of Guidance, Control, and Dynamics, Vol. 10, No. 4, 1987, pp. 401–405. doi:https://doi.org/10.2514/3.20231 JGCDDT 0162-3192 LinkGoogle Scholar[3] Peck M., "Practicable Gyroelastic Technology," 27th Annual AAS Guidance and Control Conference, AAS Paper 2004-023, Feb. 2004. Google Scholar[4] Damaren C. J. and D'Eleuterio G. M. T., "Controllability and Observability of Gyroelastic Vehicles," Journal of Guidance, Control, and Dynamics, Vol. 14, No. 5, 1991, pp. 886–894. doi:https://doi.org/10.2514/3.20728 JGCDDT 0162-3192 LinkGoogle Scholar[5] Damaren C. J. and D'Eleuterio G. M. T., "Optimal Control of Large Space Structures Using Distributed Gyricity," Journal of Guidance, Control, and Dynamics, Vol. 12, No. 5, 1989, pp. 723–731. JGCDDT 0162-3192 LinkGoogle Scholar[6] Yang L. F., Mikulas M. M. and Park K. C., "Slewing Maneuvers and Vibration Control of Space Structures by Feedforward/Feedback Moment-Gyro Controls," Journal of Dynamic Systems, Measurement, and Control, Vol. 117, No. 3, 1995, pp. 343–351. doi:https://doi.org/10.1115/1.2799125 JDSMAA 0022-0434 CrossrefGoogle Scholar[7] Shi J. F. and Damaren C. J., "Control Law for Active Structural Damping Using a Control Moment Gyro," Journal of Guidance, Control, and Dynamics, Vol. 28, No. 3, 2005, pp. 550–553. doi:https://doi.org/10.2514/1.11269 JGCDDT 0162-3192 LinkGoogle Scholar[8] Sobel K., Kaufman H. and Mabus L., "Adaptive Control for a Class of MIMO Systems," IEEE Transactions on Aerospace, Vol. 18, No. 2, 1982, pp. 576–590. CrossrefGoogle Scholar[9] Barkana I., Kaufman H. and Balas M., "Model Reference Adaptive Control of Large Structural Systems," Journal of Guidance, Control, and Dynamics, Vol. 6, No. 2, 1983, pp. 112–118. JGCDDT 0162-3192 LinkGoogle Scholar[10] Balas M., "Direct Model Reference Adaptive Control in Infinite-Dimensional Linear Spaces," Journal of Mathematical Analysis and Applications, Vol. 196, No. 1, 1995, pp. 153–171. doi:https://doi.org/10.1006/jmaa.1995.1403 JMANAK 0022-247X CrossrefGoogle Scholar[11] Barkana I. and Ben-Asher J. Z., "Simple Adaptive Control Applications to Large Flexible Structures," Journal of Guidance, Control, and Dynamics, Vol. 34, No. 6, 2011, pp. 1929–1932. doi:https://doi.org/10.2514/1.54217 JGCDDT 0162-3192 LinkGoogle Scholar[12] Livneh R. and Slater G. L., "Some Conditions for Strictly Positive Real and Almost Strictly Positive Real Structures," IEEE International Conference on Systems Engineering, Vol. 1, IEEE Publications, Piscataway, NJ, 1990, pp. 81–85. Google Scholar[13] Mehiel E. A. and Balas M. J., "Determining Strictly Positive Realness from System Modal Characteristics," Journal of Guidance, Control, and Dynamics, Vol. 30, No. 5, 2007, pp. 1539–1542. doi:https://doi.org/10.2514/1.28240 JGCDDT 0162-3192 LinkGoogle Scholar[14] Hu Q., Jia Y. and Xu S., "Recursive Dynamics Algorithm for Multibody Systems with Variable-Speed Control Moment Gyroscopes" Journal of Guidance, Control, and Dynamics, Vol. 35, No. 5, 2013, pp. 1388–1398. JGCDDT 0162-3192 LinkGoogle Scholar[15] Damaren C. J., "Optimal Control Formulation for Lightly Damped, Gyroelastic Continua," M.S. Thesis, Inst. for Aerospace Studies, Univ. of Toronto, Downsview, ON, Canada, 1987. Google Scholar[16] Wen J. T., "Time Domain and Frequency Domain Conditions for Strict Positive Realness," IEEE Transactions on Automatic Control, Vol. 33, No. 10, 1988, pp. 988–992. doi:https://doi.org/10.1109/9.7263 IETAA9 0018-9286 CrossrefGoogle Scholar[17] Anderson B. D. O. and Vongpanitlerd S., Network Analysis and Synthesis, Prentice–Hall, Englewood Cliffs, NJ, 1973, pp. 215–220. Google Scholar[18] Broussard J. R. and O'Brien M. J., "Feedforward Control to Track the Output of a Forced Model," IEEE Transactions on Automatic Control, Vol. 25, No. 4, 1980, pp. 851–853. doi:https://doi.org/10.1109/TAC.1980.1102409 IETAA9 0018-9286 CrossrefGoogle Scholar Previous article Next article

Since Almost strictly positive real (ASPR) condition is satisfied using the parallel feedforward compensator (PFC), the stable control system is easily designed with simple adaptive control (SAC) method. The augmented system output, in which the PFC is included, converges to the reference model output, and on the other hand, there is the steady-state error between the actual system output with the reference model output because of the PFC. In this study, an SAC law is extended using the null-space for controlling a rotation control system.

This paper proposes an approach of cooperative control that is based on the concept of combining neural networks and the methodology of sliding mode control (SMC). The main purpose is to eliminate the chattering phenomenon. Next, the system performance can be improved by using the method of SMC. In the present approach, two parallel Neural Networks are utilized to realize a neuro-sliding mode control (NSMC), where the equivalent control and the corrective control are the outputs of neural network 1 and neural network 2, respectively. Based on expressions of the SMC, the weight adaptations of neural network can be determined. Furthermore, the gradient descent method is used to minimize the control force so that the chattering phenomenon can be eliminated. Finally, experimental results are given to show the effectiveness and feasibility of the approach.

This paper introduces a sliding mode controller based on neural networks, for a class of single-input single-output (SISO) nonlinear discrete-time systems with unknown dynamics. The system is supposed to be represented by an input-output model implemented with layered neural networks. It is proposed a design procedure which is extended to tracking problems. It is also shown the application of this procedure to one simulated example and compared with the result from a classical controller.

This paper presents a method for face recognition based on fuzzy clustering and parallel neural networks. Neural networks have been widely used for pattern recognition in various fields. However, the computing efficiency and the recognition rate decrease rapidly if the scale of neural network increases. In this paper, a new method of face recognition based on neuro-fuzzy system is proposed. In particular, the face pattern is divided into several small-scale parallel neural networks based on fuzzy clustering. Experimental results show the effectiveness of the proposed method.

This work presents a novel integral variable structure control (IVSC) that combines a cerebellar model articulation controller (CMAC) neural network and a soft supervisor controller for use in designing single-input single-output (SISO) nonlinear system. Based on the Lyapunov theorem, the soft supervisor controller is designed to guarantee the global stability of the system. The CMAC neural network is used to perform the equivalent control on IVSC, using a real-time learning algorithm. The proposed IVSC control scheme alleviates the dependency on system parameters and eliminates the chattering of the control signal through an efficient learning scheme. The CMAC-based IVSC (CIVSC) scheme is proven to be globally stable inasmuch all signals involved are bounded and the tracking error converges to zero. A numerical simulation demonstrates the effectiveness and robustness of the proposed controller.