Multiclass Classification Fault Diagnosis of Multirotor UAVs Utilizing a Deep Neural Network

Park, Jongho; Jung, Yeondeuk; Kim, Jong-Han

doi:10.1007/s12555-021-0729-1

Cited by 21 publications

(10 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Yang et al [7] proposed a deep residual shrinkage network as a classifier using quadrotor state information (roll value, pitch value, roll rate, pitch rate, and yaw rate) and the four angular velocities of the propeller as input. Additionally, Park et al [5] used a stacked pruning sparse denoising autoencoder to process image inputs consisting of 20 signals with a length of 20 containing drone attitude, position, velocity, and acceleration information, and employed a CNN as a classifier. This approach exhibits good performance in noisy environments due to the denoising operation.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Simulation-to-reality UAV Fault Diagnosis with Deep Learning

Zhang¹,

Tong²,

Liao³

et al. 2023

Preprint

View full text Add to dashboard Cite

Accurate diagnosis of propeller faults is crucial for ensuring the safe and efficient operation of quadrotors. Training a fault classifier using simulated data and deploying it on a real quadrotor is a cost-effective and safe approach. However, the simulation-to-reality gap often leads to poor performance of the classifier when applied in real flight. In this work, we propose a deep learning model that addresses this issue by utilizing newly identified features (NIF) as input and utilizing domain adaptation techniques to reduce the simulation-to-reality gap. In addition, we introduce an adjusted simulation model that generates training data that more accurately reflects the behavior of real quadrotors. The experimental results demonstrate that our proposed approach achieves an accuracy of 96% in detecting propeller faults. To the best of our knowledge, this is the first reliable and efficient method for simulation-to-reality fault diagnosis of quadrotor propellers.

show abstract

Section: Related Workmentioning

confidence: 99%

“…In contrast, data-driven approaches, which operate in a model-free manner, have received increasing attention in recent years [4], [5]. These approaches utilize deep neural networks (DNNs) [6] to directly map sensor input to the results of the diagnosis.…”

Section: Introductionmentioning

confidence: 99%

Simulation-to-reality UAV Fault Diagnosis with Deep Learning

Zhang¹,

Tong²,

Liao³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Trained with real flight data, the classifier can perform accurate fault diagnosis on the test data collected from the same flights. Similarly, Park et al [5] proposed the stacked pruning sparse denoising auto-encoder for UAV fault diagnosis. Due to the use of denoising auto-encoder, the classifier performed well in the noisy scenario.…”

Section: A Data-driven Uav Fault Diagnosismentioning

confidence: 99%

“…Compared with model-based approaches, the model-free methods do not rely on an system model. Among modelfree approaches, data-driven approaches [4], [5] have become increasingly popular in recent years. These approaches utilize flying data and fault labels to train a fault classifier.…”

Section: Introductionmentioning

confidence: 99%

Deep Convolutional Neural Network Using Transfer Learning for Fault Diagnosis

Zhang

Zhou

2021

IEEE Access

View full text Add to dashboard Cite

Identifying the fault in propellers is important to keep quadrotors operating safely and efficiently. The simulationto-reality (sim-to-real) UAV fault diagnosis methods provide a cost-effective and safe approach to detect the propeller faults. However, due to the gap between simulation and reality, classifiers trained with simulated data usually underperform in real flights. In this work, a new deep neural network (DNN) model is presented to address the above issue. It uses the difference features extracted by deep convolutional neural networks (DD-CNN) to reduce the sim-to-real gap. Moreover, a new domain adaptation method is presented to further bring the distribution of the real-flight data closer to that of the simulation data. The experimental results show that the proposed approach can achieve an accuracy of 97.9% in detecting propeller faults in real flight. Feature visualization was performed to help better understand our DDCNN model.

show abstract

“…AFTCS is developed and implemented in hardware by Saied et al [18] for an octocopter UAV. However, a deep neural network-based fault detection algorithm is developed for the octocopter [19]. The crash probability density (CPD) is evaluated based on the Newton's laws, as well as Galileo's free fall for different types of UAVs using MATLAB simulation carried out in MATLAB in [20].To handle rotor failure in the quad-plane, a novel I-ASMC is proposed in [21].…”

Section: Introductionmentioning

confidence: 99%

Rotor Failure Compensation in a Biplane Quadrotor Based on Virtual Deflection

Dalwadi

Deb

Ožana

2022

Drones

View full text Add to dashboard Cite

A biplane quadrotor is a hybrid type of UAV that has wide applications such as payload pickup and delivery, surveillance, etc. This simulation study mainly focuses on handling the total rotor failure, and for that, we propose a control architecture that does not only handle rotor failure but is also able to navigate the biplane quadrotor to a safe place for landing. In this structure, after the detection of total rotor failure, the biplane quadrotor will imitate reallocating control signals and then perform the transition maneuver and switch to the fixed-wing mode; control signals are also reallocated. A synthetic jet actuator (SJA) is used as the redundancy that generates the desired virtual deflection to control the pitch angle, while other states are taken care of by the three rotors. The SJA has parametric nonlinearity, and to handle it, an inverse adaptive compensation scheme is applied and a closed-loop stability analysis is performed based on the Lyapunov method for the pitch subsystem. The effectiveness of the proposed control structure is validated using numerical simulation carried out in the MATLAB Simulink.

show abstract

Multiclass Classification Fault Diagnosis of Multirotor UAVs Utilizing a Deep Neural Network

Cited by 21 publications

References 45 publications

Simulation-to-reality UAV Fault Diagnosis with Deep Learning

Simulation-to-reality UAV Fault Diagnosis with Deep Learning

Deep Convolutional Neural Network Using Transfer Learning for Fault Diagnosis

Rotor Failure Compensation in a Biplane Quadrotor Based on Virtual Deflection

Contact Info

Product

Resources

About