ALFA: A dataset for UAV fault and anomaly detection

Keipour, Azarakhsh; Mousaei, Mohammadreza; Scherer, Sebastian

doi:10.1177/0278364920966642

Cited by 65 publications

(19 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…With the 19 correct sequences, our method results in 86.36% accuracy, 88.23% precision and 88.23% recall (sensitivity) over 22 flight tests. The evaluation metrics used for our calculations are explained in more details in [27].…”

Section: Resultsmentioning

confidence: 99%

“…The figures also show how the variance of the prediction error stabilizes after the model stabilization. While this work and [27] provide a dataset suitable for benchmarking different methods, to the best of our knowledge, there has been no benchmark dataset available prior to this work to enable direct comparison with the published results of similar works like Venkataraman et al [28] and Bauer et al [29]. The use of our dataset in the future will enable the comparison of methods to the state-of-the-art.…”

Section: Resultsmentioning

confidence: 99%

“…It comprises full information containing ground truth (the time and type of the failure) from 22 flight tests on a fixed-wing AAV. The data from this work is further integrated into the AIR Lab Failure and Anomaly (ALFA) Dataset, which is presented in [27] and is available at http: //theairlab.org/alfa-dataset. We modified the Ardupilot firmware to publish commanded roll and pitch signals, then compute roll error and pitch error and use these four signals as input to our method.…”

Section: B Tests and Datasetmentioning

confidence: 99%

See 2 more Smart Citations

Automatic Real-time Anomaly Detection for Autonomous Aerial Vehicles

Keipour

Mousaei

Scherer

2019

2019 International Conference on Robotics and Automation (ICRA)

Self Cite

View full text Add to dashboard Cite

The recent increase in the use of aerial vehicles raises concerns about the safety and reliability of autonomous operations. There is a growing need for methods to monitor the status of these aircraft and report any faults and anomalies to the safety pilot or to the autopilot to deal with the emergency situation. In this paper, we present a real-time approach using the Recursive Least Squares method to detect anomalies in the behavior of an aircraft. The method models the relationship between correlated input-output pairs online and uses the model to detect the anomalies. The result is an easy-to-deploy anomaly detection method that does not assume a specific aircraft model and can detect many types of faults and anomalies in a wide range of autonomous aircraft. The experiments on this method show a precision of 88.23%, recall of 88.23% and 86.36% accuracy for over 22 flight tests. The other contribution is providing a new fault detection open dataset for autonomous aircraft, which contains complete data and the ground truth for 22 fixed-wing flights with eight different types of mid-flight actuator failures to help future fault detection research for aircraft.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: B Tests and Datasetmentioning

confidence: 99%

See 1 more Smart Citation

Automatic Real-time Anomaly Detection for Autonomous Aerial Vehicles

Keipour

Mousaei

Scherer

2019

2019 International Conference on Robotics and Automation (ICRA)

Self Cite

View full text Add to dashboard Cite

show abstract

“…Throughout the study, we utilized the ALFA dataset to recognize faults when the UAV still flies, but included a fault on the system body. Note that [ 24 ] describes the detailed description of the ALFA dataset.…”

Section: Proposed Methodologymentioning

confidence: 99%

Unsupervised Fault Detection on Unmanned Aerial Vehicles: Encoding and Thresholding Approach

Park

Kim

2021

Sensors

View full text Add to dashboard Cite

Unmanned Aerial Vehicles are expected to create enormous benefits to society, but there are safety concerns in recognizing faults at the vehicle’s control component. Prior studies proposed various fault detection approaches leveraging heuristics-based rules and supervised learning-based models, but there were several drawbacks. The rule-based approaches required an engineer to update the rules on every type of fault, and the supervised learning-based approaches necessitated the acquisition of a finely-labeled training dataset. Moreover, both prior approaches commonly include a limit that the detection model can identify the trained type of faults only, but fail to recognize the unseen type of faults. In pursuit of resolving the aforementioned drawbacks, we proposed a fault detection model utilizing a stacked autoencoder that lies under unsupervised learning. The autoencoder was trained with data from safe UAV states, and its reconstruction loss was examined to distinguish the safe states and faulty states. The key contributions of our study are, as follows. First, we presented a series of analyses to extract essential features from raw UAV flight logs. Second, we designed a fault detection model consisting of the stacked autoencoder and the classifier. Lastly, we validated our approach’s fault detection performance with two datasets consisting of different types of UAV faults.

show abstract

“…Asimismo, también es posible que se pueda complementar al algoritmo con nuevas variables que pudieran ser importantes, por ejemplo, la corriente demandada por los rotores o la temperatura en cada rotor. La metodología propuesta puede aplicarse a diferentes tipos de vehículos aéreos a través de plataformas experimentales o bases de datos como las propuestas en Keipour et al (2019). Finalmente, es importante destacar que el método propuesto puede considerarse como una herramienta perfecta para poder obtener un modelo experimental mucho más preciso, del sistema a emplear.…”

Section: Conclusionesunclassified

Detección de fallas en vehículos aéreos no tripulados mediante señales de orientación y técnicas de aprendizaje de máquina

López‐Estrada

Méndez-López

Santos‐Ruiz

et al. 2021

Rev. iberoam. autom. inform. ind.

View full text Add to dashboard Cite

<p>Este trabajo propone un esquema de detección y localización de fallas en los actuadores de un vehículo aéreo no tripulado (VANT) del tipo cuadrirrotor. Para ello, se considera un enfoque basado en datos haciendo uso de técnicas de aprendizaje de máquina. En este enfoque se construye un modelo implícito del sistema a través de la información proporcionada por los sensores del VANT. Primero, a través de un plataforma de vuelo de tipo giroscópica, se captan las vibraciones correspondientes a la orientación, posición angular y aceleración lineal cuando el vehículo se encuentra en vuelo estacionario en condiciones nominales. Estos datos se procesan mediante Análisis en Componentes Principales (PCA) para la extracción de características. Posteriormente, se induce una falla a los actuadores a través de un recorte en cada una de las hélices del VANT que ocasionan una reducción del empuje generado por los rotores. Estos datos se proyectan también al subespacio de componentes principales y se comparan con los datos nominales. Para discernir entre los datos nominales y los datos cuando el vehículo presenta falla, se emplea el estadístico T2 de Hotelling. Finalmente, el desarrollo se complementa con los algoritmos de clasificación de k-vecinos más cercanos (k-NN) y de máquina de vectores de soporte (SVM). Los resultados muestran una tasa de clasificación correcta del 89.6 % (k-NN) y 92.4 %(SVM) respectivamente para 423 conjuntos de datos de validación.</p>

show abstract

ALFA: A dataset for UAV fault and anomaly detection

Cited by 65 publications

References 10 publications

Automatic Real-time Anomaly Detection for Autonomous Aerial Vehicles

Automatic Real-time Anomaly Detection for Autonomous Aerial Vehicles

Unsupervised Fault Detection on Unmanned Aerial Vehicles: Encoding and Thresholding Approach

Detección de fallas en vehículos aéreos no tripulados mediante señales de orientación y técnicas de aprendizaje de máquina

Contact Info

Product

Resources

About