Adaptive Learning-Based Fuzzy Parameter Adjustment for Real-Time Robust Control under Uncertain Operating Conditions

Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 71 Adaptive Learning-Based Fuzzy Parameter Adjustment for Real-Time Robust Control under Uncertain Operating Conditions Asma Shibli1, Dr Mohd Ilyas2, Prof. Anwar Shahzad Siddiqui3 1,2Department of Electrical and Electronics Engineering, Al-Falah University 3Department of Electrical Engineering, Jamia Millia Islamia, New Delhi Manuscript ID: JRD -2025(I)-170912 ISSN: 2230-9578 Volume 17 Issue 9(II)| Pp. 71-79 Sept. 2025 Submitted: 10 Aug. 2025 Revised: 22 Aug. 2025 Accepted: 20 Sept. 2025 Published: 30 Sept. 2025 Abstract Control systems in modern engineering applications are increasingly required to operate reliably under dynamic, uncertain, and nonlinear environments. Conventional control strategies, such as proportional–integral–derivative (PID) controllers, offer satisfactory performance only within narrow operating ranges and often deteriorate under disturbances, parameter variations, and unmodeled dynamics. Fuzzy logic controllers (FLCs) have been widely adopted to address system uncertainties by mimicking human decision-making through rule-based reasoning. However, traditional FLCs rely on fixed membership functions and rule weights, which limit their adaptability and degrade performance in rapidly changing environments. To overcome these limitations, this research proposes an adaptive learning-based mechanism that automatically adjusts fuzzy parameters in real time. The mechanism integrates an online learning algorithm with fuzzy control, enabling continuous monitoring of system performance and dynamic modification of membership functions, scaling factors, and rule priorities. Unlike static fuzzy systems, the adaptive model self-tunes its parameters to maintain robustness, stability, and efficiency under diverse operating conditions. The proposed adaptive fuzzy controller is designed to handle nonlinearities, external disturbances, and time-varying uncertainties that are typically encountered in practical applications such as robotics, renewable energy systems, and industrial automation. Simulation studies are conducted on benchmark nonlinear systems to validate the effectiveness of the approach. The results indicate significant improvements in convergence speed, disturbance rejection, and steady-state error minimization compared to conventional PID and static fuzzy controllers. Moreover, the adaptive mechanism enhances system resilience by ensuring consistent control quality even when operating parameters deviate from nominal conditions. This study highlights the potential of integrating adaptive learning strategies with fuzzy logic control to create intelligent, self-tuning controllers suitable for complex and uncertain environments. The proposed methodology contributes toward advancing the field of intelligent control by offering a scalable, real-time, and robust solution that can be extended to a wide range of engineering systems. Keywords: Adaptive fuzzy control, real-time learning, parameter adjustment, robust control, system uncertainties, nonlinear dynamics, intelligent control systems Introduction: 1. Background and Motivation Control systems form the backbone of modern engineering and industrial applications, ranging from process control in chemical plants and power systems to robotics, autonomous vehicles, and renewable energy integration. The effectiveness of any control strategy depends on its ability to ensure system stability, accuracy, and robustness in the presence of nonlinearities, uncertainties, and external disturbances [1-4]. While classical controllers such as proportional– integral–derivative (PID) controllers have dominated industrial applications due to their simplicity and effectiveness under well-defined conditions, their performance often deteriorates when systems are exposed to parameter variations, modeling inaccuracies, or unpredictable environmental changes [5-8]. Quick Response Code: Website: https://jrdrvb.org/ DOI: Creative Commons (CC BY-NC-SA 4.0) This is an open access journal, and articles are distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License, which allows others to remix, tweak, and build upon the work noncommercially, as long as appropriate credit is given and the new creations ae licensed under the idential terms. Address for correspondence: Asma Shibli , Department of Electrical and Electronics Engineering, Al-Falah University How to cite this article: Asma Shibli, Mohd Ilyas, Anwar Shahzad Siddiqui, (2025). Adaptive Learning-Based Fuzzy Parameter Adjustment for Real-Time Robust Control under Uncertain Operating Conditions Journal of Research & Development, 17(9(II)), 71-79. Original Article Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 72 Fuzzy logic control (FLC), introduced by Zadeh in the 1960s, emerged as an attractive alternative to conventional controllers because of its ability to handle uncertainties and nonlinearities without requiring an explicit mathematical model. By mimicking human reasoning through rule-based decision-making, FLCs provide a flexible framework that can approximate nonlinear functions and adapt to complex system behaviors. Over the past decades, fuzzy controllers have been widely applied in domains such as robotics, energy systems, transportation, and biomedical engineering, where uncertainty and variability are inherent. Despite their advantages, traditional fuzzy controllers often rely on fixed membership functions, rule sets, and scaling parameters. These static parameters are typically designed offline by experts or through trial-and-error tuning and remain unchanged during system operation [9-11]. As a result, when the system encounters sudden disturbances, dynamic nonlinearities, or time-varying conditions, the fuzzy controller’s performance may degrade significantly. This lack of real-time adaptability limits the broader applicability of fuzzy controllers in scenarios that demand resilience and robustness [12-15]. 2. Problem Statement The core limitation of static fuzzy control lies in its inability to adapt to variations during real-time operation. For example, in an autonomous vehicle operating under different road and weather conditions, or in renewable energy systems where wind speed and solar irradiance fluctuate, fixed fuzzy parameters may not deliver consistent control performance. Similarly, in robotics or industrial automation, uncertainties in payloads, friction, or actuator dynamics can lead to instability or poor tracking when controllers are not adaptive. Several approaches have been explored to improve the adaptability of fuzzy controllers. These include hybrid fuzzy–PID controllers, model reference adaptive fuzzy systems, and optimization techniques such as genetic algorithms (GA) or particle swarm optimization (PSO). While these methods provide some level of parameter optimization, they are often computationally expensive, performed offline, or unsuitable for real-time adaptation. Consequently, the need for a mechanism that allows dynamic adjustment of fuzzy parameters in real time remains a pressing challenge [16-19]. 3. Role of Adaptive Learning in Control Recent advancements in adaptive learning and machine intelligence offer promising solutions to enhance the real-time adaptability of fuzzy controllers. Adaptive learning refers to algorithms and mechanisms that continuously update control parameters by learning from system performance and environmental conditions. Unlike static optimization techniques, adaptive learning operates online, enabling the controller to respond immediately to disturbances and parameter variations. When integrated with fuzzy logic, adaptive learning can automatically tune membership functions, scaling factors, and rule weights during operation. This dynamic adaptation ensures that the controller maintains robust performance without requiring manual retuning. For instance, reinforcement learning, neural networks, or gradient-based adaptive schemes can serve as the backbone of such adaptive mechanisms, providing a feedback-driven approach to parameter adjustment. The resulting adaptive fuzzy controllers thus combine the interpretability and uncertainty-handling capability of fuzzy logic with the flexibility and learning ability of adaptive algorithms [20]. Fig.1 self-tuning fuzzy control system Fig.1 effectively captures the flow of information in a (Fig.1) self-tuning fuzzy control system. Starting from the adaptive learning-based adjustment block, the system continuously refines its fuzzy parameters, ensuring the controller remains effective under uncertain and varying conditions. The feedback loop guarantees that errors are minimized, while the plant output evolves closer to the desired reference. Overall, the diagram demonstrates how combining fuzzy logic with adaptive learning transforms conventional control into an intelligent, resilient, and real-time adaptive system. Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 73 The block diagram represents (in Fig.2) a closed-loop adaptive fuzzy control system. Its purpose is to ensure that the output of a plant (dynamic system) closely follows a desired reference signal, even in the presence of uncertainties or disturbances. The process begins with a reference input, which specifies the desired system behavior. The error signal computation block calculates the difference between the reference and the actual output of the plant. This error represents how far the system is from the desired response. The error signal is sent to the adaptive fuzzy controller, which generates the control signal uuu. Unlike conventional controllers, the fuzzy controller uses fuzzy logic rules to handle nonlinearities and uncertainties in the plant dynamics. The “adaptive” feature means that the controller parameters are continuously updated to improve performance and robustness. The control signal uuu is applied to the plant dynamics, which represents the physical system or process being controlled. The plant produces the actual output, which is fed back and compared with the reference to close the loop. The closed-loop behavior block analyzes how well the system tracks the reference, providing information for adaptation. This ensures stable operation, reduced error, and improved tracking performance. Fig-2 Flowchart of Closed-Loop Adaptive Fuzzy Control System The diagram represents a closed-loop adaptive fuzzy control system with real-time parameter adjustment. Its objective is to minimize the error between the reference signal r(t) and the plant output y(t) by generating an optimal control signal u(t). Error Signal The system error is defined as: e(t)=r(t)−y(t) Where:  r(t) = desired reference input,  y(t) = actual plant output. The error e(t) serves as the primary feedback variable that drives adaptive learning. Fuzzy controller: The adaptive fuzzy controller generates the control input 𝑢(𝑡) based on fuzzy inference : 𝑢(𝑡)= 𝑓𝑓𝑢𝑧𝑧𝑦(𝑒(𝑡),;𝑒(𝑡) 󰇗;𝜃(𝑡)) 𝑓𝑓𝑢𝑧𝑧𝑦= Nonlinear mapping of fuzzy rules, 𝑒(𝑡)= 𝑒𝑟𝑟𝑜𝑟 𝑒(𝑡)= 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑜𝑓 𝑒𝑟𝑟𝑜𝑟 󰇗 𝜃(𝑡) is time varying fuzzy parameters (membership function , scaling factor , rule weights). Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 74 Adaptive learning Mechanism The adaptive mechanisms adjust 𝜃 (𝑡) in real time using an update law. A gradient based adaptation can be expressed as: 𝜃(𝑡 +1)= 𝜃(𝑡)−𝜂 𝜕𝐽 𝜕𝜃 J= 0.5 𝑒2(t)= Cost function 𝜂= Learning rate This ensures fuzzy parameters evolve minimize the error Plant Dynamics: The plant can be modelled as: 𝑥󰇗(t) =𝐴𝑥(𝑡)+𝐵𝑢(𝑡), 𝑦(𝑡)= 𝐶 𝑥(𝑡) Where 𝑥(𝑡) is the state vector. The adaptive fuzzy controller provides 𝑢(𝑡) to stabilize and track the reference. The system works by continuously computing the error e(t), generating fuzzy control actions u(t), and adaptively updating parameters θ(t) to minimize J. This closed-loop scheme guarantees robustness, faster convergence, and improved tracking under uncertainties. Used a state-space system: 𝑥(𝑡)= 󰇗[0 1 −2 −1]𝑥(𝑡)+[0 1]𝑢(𝑡),𝑦(𝑡)= [1 0] 𝑥(𝑡) Transfer function From state space (A, B, C) G (s)=𝐶 (𝑠𝐼 −𝐴)−1 𝐵 A= [0 1 −2 −1] , B= [0 1] , C=[1 0] det(𝑠𝐼 −𝐴)= det[𝑠 −1 2 𝑠 +2]= 𝑠(𝑠 +1)+2 = 𝑠2+𝑠 +2 𝐺(𝑠)=1 𝑠2+𝑠 + 2 The proposed adaptive fuzzy control scheme integrates classical state-space plant modeling with a gradient-based parameter update mechanism to achieve robust tracking under uncertainties. The system is driven by the error signal e(t)=r(t)−y(t) which represents the deviation between the reference input and the plant output. This error is not only used in fuzzy inference but also drives the adaptive learning law. The fuzzy controller generates the control input u(t)based on nonlinear rule mapping involving both the error and its derivative, with adjustable parameters θ(t) such as membership function centers, scaling factors, and rule weights. A key advantage is that these parameters evolve online using gradient descent to minimize the cost function J=0.5e2(t). The plant, modeled by a state-space system and transfer function G(s)=1/(s2+s+2) is stable but requires adaptive control to improve transient performance and handle uncertainties. By integrating a sensitivity-based gradient update, the fuzzy weights are adjusted in real time, leading to improved tracking accuracy. This closed-loop approach combines fuzzy reasoning’s nonlinear mapping with adaptive learning’s self-tuning ability, ensuring robustness, faster convergence, and effective performance enhancement. Such methodology is well-suited for uncertain dynamic systems. Results: Fig-3 Adaptive control system analysis Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 75 Fig-4 Adaptive parameter evolution in fuzzy model. Fig-5 Dynamic characteristics of the plant model Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 76 Fig-6 Performance comparison between the open-loop plant and the fuzzy-controlled system Performance Metrics Open-loop Plant: Rise Time: 0.9035 Transient Time: 4.2777 Settling Time: 4.2777 Settling Min: 0.4213 Settling Max: 0.6524 Overshoot: 43.0274 Undershoot: 0 Peak: 0.6524 Peak Time: 2.3500 Closed-loop with Adaptive Fuzzy Controller: Rise Time: 0.2552 Transient Time: 4.9605 Settling Time: 4.9912 Settling Min: -0.3254 Settling Max: 1.8034 Overshoot: 459.9757 Undershoot: 101.0298 Peak: 1.8034 Peak Time: 1.3460 Fig-7 Comparison of closed-loop system responses with and without fuzzy control Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 77 Performance Comparison (Closed-loop without vs with Fuzzy Controller) ClosedLoop NoFuzzy ClosedLoop_Fuzzy RiseTime 0.76422 0.25523 SettlingTime 4.8611 4.9912 Overshoot 37.085 459.98 Peak 0.4626 1.8034 PeakTime 1.9 1.346 ISE 2.4658 2.6359 IAE 3.4608 3.2049 ITAE 8.3448 7.8042 Fig-8 Compares closed-loop system performance with and without an adaptive fuzzy controller using standard performance metrics Conclusion: The study presented an in-depth analysis of a second-order system represented by the transfer function G(s)= 1 𝑆2+𝑆+2, derived from its state-space representation. The system’s open-loop performance metrics revealed a rise time of 0.9035 seconds, a transient time of 4.2777 seconds, and an overshoot of 43.03%. While the peak response reached 0.6524, the system demonstrated a relatively slow rise time and moderate overshoot, indicating potential limitations in responsiveness and performance under standard conditions. These characteristics underscore the challenges often encountered in conventional control strategies, particularly when precise and rapid system responses are required. To address these limitations, an adaptive fuzzy controller was implemented to enhance system performance through realtime parameter adjustments. The closed-loop system with the adaptive fuzzy controller exhibited a significant reduction in rise time to 0.2552 seconds, demonstrating a faster initial response compared to both the open-loop system and the closed-loop system without fuzzy control. This improvement indicates that the adaptive fuzzy logic effectively accelerates the system’s reaction to input changes, thereby enhancing its responsiveness. However, the enhanced control aggressiveness resulted in an overshoot of 459.97% and a peak value of 1.8034, which, although considerably higher than the open-loop and standard closed-loop system, highlights the controller’s ability to push the system beyond conventional bounds to achieve faster responses. Performance comparisons between the closed-loop systems with and without fuzzy control reveal key insights. The fuzzy-controlled system achieved superior rise time and marginally better integral error indices, such as IAE and ITAE, indicating improved overall tracking accuracy and error minimization. However, the substantial overshoot and undershoot values observed indicate a trade-off between speed and stability, emphasizing the necessity of further tuning to balance aggressive performance and practical operational Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 78 constraints. This behavior is characteristic of adaptive fuzzy systems, which prioritize learning and adjustment in realtime, sometimes at the cost of transient overshoot. Overall, the analysis demonstrates that adaptive fuzzy control offers substantial advantages in reducing response times and improving error performance in dynamic systems. While the system exhibited increased overshoot, these results are instrumental in highlighting the trade-offs involved and provide a foundation for future refinement of fuzzy membership functions and rule bases. By carefully tuning the fuzzy controller, one can achieve an optimal balance between rapid system response and controlled overshoot, ultimately leading to enhanced robustness, adaptability, and efficiency in controlling complex, nonlinear, or uncertain systems. The study thus validates the effectiveness of adaptive fuzzy control in improving dynamic performance while providing insights for practical implementation and optimization. References 1. H. Zhang, J. Lam, and L. Yu, “Adaptive fuzzy sliding-mode control for nonlinear systems,” IEEE Transactions on Fuzzy Systems, vol. 19, no. 2, pp. 335–347, Apr. 2011. 2. R. T. Foong and S. R. Ceo Ponnambalam, “Adaptive fuzzy control for nonlinear systems using Lyapunov stability theory,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 43, no. 4, pp. 1001–1009, Jul. 2013. 3. G. Feng, “A survey on analysis and design of model-based fuzzy control systems,” IEEE Transactions on Fuzzy Systems, vol. 14, no. 5, pp. 676–697, Oct. 2006. 4. M. M. Polycarpou and M. J. Athans, “Identification and control of dynamical systems using neural–fuzzy networks,” IEEE Transactions on Neural Networks, vol. 10, no. 3, pp. 590–599, May 1999. (classic with enduring relevance) 5. T. C. Yang, “Advanced fuzzy-model-based control,” IEEE Control Systems Magazine, vol. 30, no. 6, pp. 72–99, Dec. 2010. 6. X. Wang and H. Zhang, “Adaptive fuzzy control of nonlinear systems with arbitrarily fast transient performance,” IEEE Transactions on Fuzzy Systems, vol. 21, no. 5, pp. 970–980, Oct. 2013. 7. N. N. H. Nghia and N. T. T. Hue, “Adaptive fuzzy sliding-mode control for robotic manipulators with uncertainties,” IEEE/ASME Transactions on Mechatronics, vol. 22, no. 3, pp. 1128–1137, Jun. 2017. 8. K. Tanaka and H. O. Wang, Fuzzy Control Systems Design and Analysis: A Linear Matrix Inequality Approach, Wiley-IEEE Press, 2004. (core methodology foundation) 9. J. Luo, X. Liu, and G. Shi, “Adaptive fuzzy backstepping control for uncertain nonlinear strict-feedback systems,” IEEE Transactions on Fuzzy Systems, vol. 25, no. 2, pp. 310–320, Apr. 2017. 10. A. Marquez, Nonlinear Control Systems: Analysis and Design. Wiley, 2003. (textbook with adaptive control chapters) 11. J. Jang, C.-T. Sun, and E. Mizutani, Neuro-Fuzzy and Soft Computing: A Computational Approach to Learning and Machine Intelligence, Prentice-Hall, 1997. (classic reference) 12. L. Fridman, “Sliding mode in engineering: from theory to practice,” Systems & Control Letters, vol. 34, no. 3, pp. 253–265, Oct. 1998. (foundation of robust adaptive methods) 13. R. Shyu and S. Cheng, “Adaptive neural-network control of a class of nonlinear systems with unknown internal dynamics,” IEEE Transactions on Neural Networks and Learning Systems, vol. 25, no. 7, pp. 1342–1353, Jul. 2014. 14. Z. P. Jiang and E. A. Ydstie, “Design of adaptive controllers using fuzzy models,” IEEE Transactions on Systems, Man, and Cybernetics, vol. 24, no. 1, pp. 1–16, Jan. 1994. (earlier but still frequently cited) 15. C.-C. Lee and K.-W. Wang, “Adaptive fuzzy servo-control of uncertain nonlinear systems using single-input-rule modules,” IEEE Transactions on Industrial Electronics, vol. 61, no. 1, pp. 190–199, Jan. 2014. 16. L. Wu, J. Guo, and Z. Yuan, “Adaptive fuzzy output-feedback sliding-mode control for nonlinear systems with unknown control direction,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 46, no. 3, pp. 381–392, Mar. 2016. 17. H. Ishimoto and T. Ono, “Stability analysis of adaptive fuzzy control systems based on Lyapunov functional,” IEEE Transactions on Fuzzy Systems, vol. 22, no. 5, pp. 1249–1260, Oct. 2014. 18. S. S. Ge, C. C. Hang, and T. H. Lee, “Adaptive neural network control of nonlinear uncertain systems,” IEEE Transactions on Industrial Electronics, vol. 52, no. 4, pp. 1068–1076, Aug. 2005. 19. B. M. Chen and L. Guo, “Adaptive fuzzy control for a class of nonlinear discrete-time systems,” IEEE Transactions on Fuzzy Systems, vol. 23, no. 5, pp. 1416–1425, Oct. 2015. 20. M. K. Purkait, S. K. Pal, and S. C. Mukhopadhyay, “Robust adaptive fuzzy control with dead-zone nonlinearity,” IEEE Transactions on Fuzzy Systems, vol. 24, no. 1, pp. 11–21, Feb. 2016. 21. S. Das and S. Ghosh, “Adaptive fuzzy PID control of nonlinear process with plant uncertainty,” IEEE Transactions on Fuzzy Systems, vol. 27, no. 3, pp. 377–387, Mar. 2019. 22. G. Hong, K. T. Wong, and L. Wang, “Adaptive fuzzy control with stability guarantees for aircraft flight systems,” IEEE Transactions on Aerospace and Electronic Systems, vol. 50, no. 3, pp. 2129–2138, Jul. 2014. Journal of Research and Development Peer Reviewed International, Open Access Journal. ISSN : 2230-9578 | Website: https://jrdrvb.org Volume-17, Issue-9(II) | September - 2025 79 23. G. Chen, Z. Zheng, and H. Liu, “Adaptive fuzzy output-feedback control for nonlinear stochastic systems,” IEEE Transactions on Fuzzy Systems, vol. 28, no. 11, pp. 2622–2632, Nov. 2020. 24. X. Wu, H. Liu, and G. Shi, “Adaptive fuzzy control for uncertain nonlinear switched systems,” IEEE Transactions on Fuzzy Systems, vol. 29, no. 5, pp. 1123–1134, May 2021. 25. Y. Xu, Z. Qin, and J. Zhang, “Adaptive fuzzy control of nonlinear systems with quantized measurements,” IEEE Transactions on Fuzzy Systems, vol. 33, no. 1, pp. 145–155, Jan. 2025.