TL;DR
This paper introduces a novel adaptive fuzzy controller for Flapping Wing Micro Air Vehicles (FW MAVs) that leverages Sliding Mode Control (SMC) theory for robust altitude control. By dynamically tuning fuzzy system parameters, the controller demonstrates superior performance and adaptability compared to traditional PID controllers in various challenging flight scenarios.
What Problem Does This Paper Address?
Controlling Flapping Wing Micro Air Vehicles (FW MAVs) is a significant challenge due to their inherent nonlinearities, over-actuation, and susceptibility to environmental disturbances. Traditional control methods often require precise mathematical models, which are difficult to obtain for these complex systems. This paper addresses the need for a robust, model-free adaptive control strategy that can effectively manage the altitude of FW MAVs, enabling them to perform maneuvers like vertical take-off and landing, hovering, and fast turns.
Methods
The authors propose an adaptive fuzzy controller that combines the strengths of Fuzzy Logic Control (FLC) and Sliding Mode Control (SMC). The controller utilizes a Takagi-Sugeno (T-S) fuzzy structure. The antecedent parameters (centers and widths of membership functions) are tuned using Fuzzy C-Means (FCM) clustering, with the error signal serving as the cost function. Crucially, the consequent parameters of the fuzzy system are adapted using SMC theory-based adaptation laws, which also guarantee the closed-loop stability of the controller. The controller's output is designed to adjust the flapping amplitude of the FW MAV, as this parameter was identified as the dominant factor for altitude control.
Datasets
The paper does not explicitly mention training datasets. The evaluation was performed using a simulated FW MAV model. The performance of the proposed controller was compared against a PID controller using several reference trajectories:
- Constant height of 10 m.
- Sinusoidal wave function with an amplitude of 1 m and frequency of 1 Hz.
- Two different step functions: one changing altitude from 0 m to 10 m after 20 seconds, and another changing from 5 m to 10 m after 20 seconds.
Key Insights and Ideas
- Synergistic Control: Combining FCM for antecedent tuning and SMC for consequent adaptation offers a powerful approach to adaptive fuzzy control, addressing both parameter uncertainty and stability guarantees.
- Model-Free Adaptability: The controller's ability to adapt to environmental disturbances without relying on an exact FW MAV model is a significant advantage for real-world applications.
- Flapping Amplitude as Key Control Input: Identifying flapping amplitude as the primary actuator for altitude control simplifies the control problem and allows for focused adaptation.
- SMC for Stability: The use of SMC theory not only drives adaptation but also provides a theoretical foundation for ensuring the stability of the closed-loop system, a critical aspect for aerial vehicles.
Key Results
The proposed adaptive SMC-Fuzzy controller consistently outperformed a conventional PID controller across all tested scenarios. Specifically:
- For a constant height trajectory, the SMC-Fuzzy controller showed better performance with less overshoot.
- Similar improvements were observed for the sinusoidal reference signal.
- In step function tests, the SMC-Fuzzy controller demonstrated satisfactory and superior performance compared to the PID controller.
- The Root Mean Square Error (RMSE) values in Table 1 confirm the superiority of the SMC-Fuzzy controller, with consistently lower RMSEs across all reference signals compared to the PID controller. For example, the RMSE for the constant height trajectory was 0.6460 for SMC-Fuzzy versus 0.7891 for PID.
Limitations and Future Work
- Simulation-Based: The current evaluation is based on a simulated FW MAV model. The authors acknowledge that future work will involve implementing and testing the proposed controller in a real-world FW MAV flight test.
- Altitude Control Focus: While the paper focuses on altitude control, future work could extend this adaptive fuzzy controller to manage other flight dynamics (e.g., pitch, roll, yaw) for more comprehensive control.
- Parameter Tuning: While FCM and SMC are used for tuning, further investigation into the optimal initialization and interaction of these tuning mechanisms could yield even better performance.
Summary updated