Fractional-Order Adaptive Fault-Tolerant Synchronization Tracking Control of Networked Fixed-Wing UAVs Against Actuator-Sensor Faults via Intelligent Learning Mechanism

Ziquan Yu,Youmin Zhang,Bin Jiang,Chun-Yi Su,Jun Fu,Ying Jin,Tianyou Chai

DOI: https://doi.org/10.1109/tnnls.2021.3059933

IF: 14.255

2021-12-01

IEEE Transactions on Neural Networks and Learning Systems

Abstract:This article presents an enhanced fault-tolerant synchronization tracking control scheme using fractional-order (FO) calculus and intelligent learning architecture for networked fixed-wing unmanned aerial vehicles (UAVs) against actuator and sensor faults. To increase the flight safety of networked UAVs, a recurrent wavelet fuzzy neural network (RWFNN) learning system with feedback loops is first designed to compensate for the unknown terms induced by the inherent nonlinearities, unexpected actuator, and sensor faults. Then, FO sliding-mode control (FOSMC), involving the adjustable FO operators and the robustness of SMC, are dexterously proposed to further enhance flight safety and reduce synchronization tracking errors. Moreover, the dynamic parameters of the RWFNN learning system embedded in the networked fixed-wing UAVs are updated based on adaptive laws. Furthermore, the Lyapunov analysis ensures that all fixed-wing UAVs can synchronously track their references with bounded tracking errors. Finally, comparative simulations and hardware-in-the-loop experiments are conducted to demonstrate the validity of the proposed control scheme.

computer science, artificial intelligence, theory & methods,engineering, electrical & electronic, hardware & architecture

What problem does this paper attempt to address?

The paper is primarily dedicated to addressing the issue of fault-tolerant synchronized tracking control for networked fixed-wing Unmanned Aerial Vehicles (UAVs) in the presence of actuator and sensor faults. Specifically:

1. **Research Background**: With the development of UAV technology, the cooperative control of networked fixed-wing UAVs has been widely applied in many fields such as power line/pipeline inspection, surveillance and monitoring, and aerial operations. However, during flight, actuator and sensor faults can severely impact the performance of UAV swarms and even lead to catastrophic accidents.

2. **Main Objective**: To enhance the safety and reliability of networked fixed-wing UAVs, this paper proposes an enhanced fault-tolerant synchronized tracking control scheme based on fractional calculus theory and intelligent learning mechanisms.

3. **Technical Approaches**:
- A Recurrent Wavelet Fuzzy Neural Network (RWFNN) learning system with a feedback loop is designed to compensate for unknown terms caused by inherent nonlinearity and unexpected actuator and sensor faults.
- A Fractional Order Sliding Mode Control (FOSMC) is proposed, which combines the adjustable fractional order operator and the robustness of sliding mode control to further improve flight safety and reduce synchronized tracking errors.
- Adaptive laws are utilized to update the dynamic parameters of the RWFNN learning system embedded in the networked fixed-wing UAVs.
- Lyapunov analysis is employed to ensure that all fixed-wing UAVs can synchronously track within their reference signal range and that the tracking error is bounded.

4. **Validation Method**: The effectiveness of the proposed control scheme is validated through comparative simulations and hardware-in-the-loop experiments.

5. **Contributions**:
- The paper investigates the fault-tolerant synchronized tracking control problem for networked fixed-wing UAVs in the presence of actuator and sensor faults, which is rarely addressed in existing research.
- For the first time, a Recurrent Wavelet Fuzzy Neural Network learning system is applied as a fault-tolerant control unit in fractional order sliding mode synchronized tracking control.
- A fractional order sliding mode error is constructed to enhance fault-tolerant control performance and increase the number of adjustable control parameters.
- The practical effectiveness of the proposed scheme is experimentally validated.

In summary, the main aim of this paper is to develop a new control strategy that ensures the safe operation of networked fixed-wing UAVs in the presence of actuator and sensor faults.