RESUMEN
As anomaly detection for electrical power steering (EPS) systems has been centralized using model- and knowledge-based approaches, EPS system have become complex and more sophisticated, thereby requiring enhanced reliability and safety. Since most current detection methods rely on prior knowledge, it is difficult to identify new or previously unknown anomalies. In this paper, we propose a deep learning approach that consists of a two-stage process using an autoencoder and long short-term memory (LSTM) to detect anomalies in EPS sensor data. First, we train our model on EPS data by employing an autoencoder to extract features and compress them into a latent representation. The compressed features are fed into the LSTM network to capture any correlated dependencies between features, which are then reconstructed as output. An anomaly score is used to detect anomalies based on the reconstruction loss of the output. The effectiveness of our proposed approach is demonstrated by collecting sample data from an experiment using an EPS test jig. The comparison results indicate that our proposed model performs better in detecting anomalies, with an accuracy of 0.99 and a higher area under the receiver operating characteristic curve than other methods providing a valuable tool for anomaly detection in EPS.
Asunto(s)
Aprendizaje Profundo , Reproducibilidad de los Resultados , Electricidad , Curva ROCRESUMEN
A wavelet Electrocardiogram (ECG) detector for low-power implantable cardiac pacemakers is presented in this paper. The proposed wavelet-based ECG detector consists of a wavelet decomposer with wavelet filter banks, a QRS complex detector of hypothesis testing with wavelet-demodulated ECG signals, and a noise detector with zero-crossing points. In order to achieve high detection accuracy with low power consumption, a multi-scaled product algorithm and soft-threshold algorithm are efficiently exploited in our ECG detector implementation. Our algorithmic and architectural level approaches have been implemented and fabricated in a standard 0.35 µm CMOS technology. The testchip including a low-power analog-to-digital converter (ADC) shows a low detection error-rate of 0.196% and low power consumption of 19.02 µW with a 3 V supply voltage.