Detecting known and unknown faults in automotive systems using ensemble-based anomaly detection

Theissler, Andreas

doi:10.1016/j.knosys.2017.02.023

Cited by 122 publications

(58 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A datadriven classifier approach for fault diagnosis of an electric throttle control system is proposed in [5] where incremental learning is used to improve classification performance over time. In [11], an ensemble approach for automotive fault classification of both known and unknown faults in time-series data is developed by combining multiple machine learning methods for classification. A two step fault classification approach to handle unknown faults in an electronic system using Gaussian mixture models and k-means is proposed in [12].…”

Section: B Related Researchmentioning

confidence: 99%

Data-Driven Open-Set Fault Classification of Residual Data Using Bayesian Filtering

Jung

2020

IEEE Trans. Contr. Syst. Technol.

View full text Add to dashboard Cite

Data-driven fault classification in industrial applications is complicated by unknown fault classes and limited training data. In addition, different faults can have similar effects on sensor outputs resulting in fault classification ambiguities, i.e. multiple fault hypotheses can explain the data. One solution is to identify and rank all plausible fault classes which give useful information, for example at a workshop when performing troubleshooting. A probabilistic fault classification algorithm is proposed for residual data classification combining Weibullcalibrated one-class support vector machines for fault class modeling and Bayesian filtering for time-series analysis. The fault classifier ranks different fault classes and can identify sequences from unknown fault realizations, i.e. faults not represented in training data. Real residual data computed from sensor data and model analysis of an internal combustion engine are used as a case study illustrating the usefulness of the proposed method.

show abstract

Section: B Related Researchmentioning

confidence: 99%

Data-Driven Open-Set Fault Classification of Residual Data Using Bayesian Filtering

Jung

2020

IEEE Trans. Contr. Syst. Technol.

View full text Add to dashboard Cite

show abstract

“…An ensemble of five different binary classifier system is proposed to discover sensor anomalies, and the weighted majority voting algorithm is used to take the ensemble final decision [26]. Theissler [27] proposed an ensemble classifier algorithm that is capable of detecting faults of known and unknown fault types in automotive systems. The diversity required for effective ensemble methods is modeled by using twoclass classifiers and one-class classifiers, and the selection of the base classifiers is done statically.…”

Section: Related Workmentioning

confidence: 99%

Anomaly Detection for Controller Area Network in Braking Control System With Dynamic Ensemble Selection

Yang

Wang

et al. 2019

IEEE Access

View full text Add to dashboard Cite

The controller area networks (CAN) in the braking control system of metro trains are used to transmit the important control instruction and condition information, whose anomaly will endanger the security of trains running seriously. Due to the harsh work environment, there are various known and previously unknown fault types, current scheduled maintenance cannot detect early anomaly in time, and constructing an accurate and stable anomaly detector is a challenging task. In this paper, an anomaly detection approach is proposed to detect anomaly based on a dynamic ensemble selection system (DESS) without the expert knowledge, which involves two-class and one-class classifiers, and the base classifiers are trained with the network features extracted from the physical-layer information. To conduct the fusion, the support function of ''distance-based'' classifier is redefined as a class-conditional probability density function, and the source competence of base classifier is estimated by the entropy-based method in validate space and extended to entire decision space using the normalized Gaussian potential function. For different fault types, the competence classifiers are selected and the anomaly detection result is finally achieved by weighted majority voting. The comparative experiments are included in this paper to demonstrate the effectiveness and robustness in anomaly detection, including varying fault types.

show abstract

“…Zhang et al (2009) also used fleets of vehicles for detecting and isolating unexpected faults in the production stage. Recently Theissler (2017) has provided the categorisation of anomalies in automotive data, and stressed the importance of designing detection methods that can handle both known and unknown fault types, together with validation on real data.…”

Section: Related Workmentioning

confidence: 99%

Self-monitoring for maintenance of vehicle fleets

Rögnvaldsson

Nowaczyk

Byttner

et al. 2017

Data Min Knowl Disc

View full text Add to dashboard Cite

An approach for intelligent monitoring of mobile cyberphysical systems is described, based on consensus among distributed self-organised agents. Its usefulness is experimentally demonstrated over a long-time case study in an example domain: a fleet of city buses. The proposed solution combines several techniques, allowing for life-long learning under computational and communication constraints. The presented work is a step towards autonomous knowledge discovery in a domain where data volumes are increasing, the complexity of systems is growing, and dedicating human experts to build fault detection and diagnostic models for all possible faults is not economically viable. The embedded, self-organised agents operate on-board the 123Self-monitoring for maintenance of vehicle fleets 345 cyberphysical systems, modelling their states and communicating them wirelessly to a back-office application. Those models are subsequently compared against each other to find systems which deviate from the consensus. In this way the group (e.g., a fleet of vehicles) is used to provide a standard, or to describe normal behaviour, together with its expected variability under particular operating conditions. The intention is to detect faults without the need for human experts to anticipate them beforehand. This can be used to build up a knowledge base that accumulates over the life-time of the systems. The approach is demonstrated using data collected during regular operation of a city bus fleet over the period of almost 4 years.

show abstract

Detecting known and unknown faults in automotive systems using ensemble-based anomaly detection

Cited by 122 publications

References 21 publications

Data-Driven Open-Set Fault Classification of Residual Data Using Bayesian Filtering

Data-Driven Open-Set Fault Classification of Residual Data Using Bayesian Filtering

Anomaly Detection for Controller Area Network in Braking Control System With Dynamic Ensemble Selection

Self-monitoring for maintenance of vehicle fleets

Contact Info

Product

Resources

About