Vibration Control of Passenger Aircraft Active Landing Gear Using Neural Network-Based Fuzzy Inference System

Analysis

Why This Paper Matters

Runway surface roughness is a major source of vibration in passenger aircraft during taxiing, affecting ride comfort, safety, and human health. While previous studies have explored passive, semi-active, and intelligent controllers like PID, H∞, and ANFIS, the comprehensive application of a robust adaptive neuro-fuzzy inference system (RANFIS) to active landing-gear control has not been addressed. This paper fills that gap by proposing RANFIS, which combines robustness with adaptive learning of fuzzy rules and neural network parameters, offering a novel approach to active vibration control.

The significance of this work lies in its potential to improve passenger experience and operational reliability. By reducing vibration peaks, settling times, and steady-state errors, RANFIS can enhance comfort and safety while also contributing to structural durability. This is particularly relevant for next-generation aircraft, where active control systems are increasingly being integrated to meet stringent performance and safety standards.

Technical Contributions

• Novel RANFIS Controller: The paper introduces a robust adaptive neuro-fuzzy inference system that integrates robustness with adaptive learning, addressing a critical gap in the literature.
• Six-DOF Aircraft Model: A comprehensive six-degrees-of-freedom dynamic model is developed to simulate aircraft behavior under stochastic runway excitations, providing a realistic testbed for controller evaluation.
• Comparative Evaluation: The study systematically compares MBNN, ANFIS, and RANFIS across multiple performance metrics, offering insights into their relative strengths and weaknesses.
• Performance Metrics: The evaluation focuses on rise time, settling time, peak value, and steady-state error, providing a thorough assessment of control quality.

Results

The simulation results demonstrate that RANFIS achieves superior performance compared to MBNN and ANFIS. Specifically, RANFIS reduces vibration peaks to ≤0.3–1.0 cm, shortens settling times to <1.5 s, and decreases steady-state errors to <0.05 cm. These metrics indicate a significant improvement in vibration suppression and control accuracy. While MBNN and ANFIS provide satisfactory control, RANFIS consistently outperforms them, confirming its effectiveness in enhancing comfort, safety, and structural durability.

Significance

The broader impact of this research extends to the field of AI-based control systems, particularly in aerospace applications. By demonstrating the effectiveness of a robust neuro-fuzzy approach, the paper encourages further exploration of hybrid intelligent controllers that combine robustness with adaptive learning. This could lead to more reliable and efficient active control systems not only in landing gear but also in other dynamic systems where vibration and disturbance rejection are critical. The findings also highlight the potential of simulation-based studies to guide the design of next-generation aircraft systems, paving the way for experimental validation and real-world deployment.