This paper proposes a wearable device-based system to monitor the abnormal conditions of a driver, including stress, fatigue, and drowsiness. The system measures the motional and physiological information of the driver using the developed wearable device on the wrist. Preprocessing is used to distinguish the valid signal parts of the measured signals, because various noises can occur in wearable sensors. Features are extracted from the signal parts, and an optimal feature set is determined by an analysis of variance and a sequential floating forward selection algorithm. To classify the driver's state, a support vector machine-based classification method is used to obtain high generalization performance considering interdriver variance. Experiments were conducted on an indoor driving simulator, with 28 subjects, to gather data for each state. The classification accuracy was 98.43% for fivefold cross validation on the data. In a subject-independent test, the accuracy was 68.31% for the four states and 84.46% for the three states consisting of normal, stressed, and fatigued or drowsy states. Using the proposed system, the abnormal conditions of the driver can be detected and distinguished. This advantage contributes to safer and more comfortable driving. Furthermore, the utilization of the wearable device makes the system easy to use.

With the development of smart factories, deep learning, which automatically extracts features and diagnoses faults, has become an important approach for fault diagnosis. In this paper, a novel interturn short-circuit fault (ISCF) diagnosis approach using an attention based recurrent neural network is proposed. An encoder-decoder architecture using an attention mechanism diagnoses the ISCF by estimating a fault indicator that directly reflects the severity of the fault, using currents and rotational speed signals as inputs. The attention mechanism helps the decoding process in accurate diagnosis and solves the long-term dependency problem of the encoder-decoder structure. The proposed algorithm uses only three-phase current and rotational speed as the inputs to evaluate the severity of the ISCF and enable early stage diagnosis of ISCF. The diagnosis of ISCF is achieved in various operating points and fault conditions, and no additional sensors such as voltage and vibration sensors are required. Experimental results for various operating and fault conditions demonstrate that the proposed method effectively diagnoses ISCFs.

As interturn short fault (ISF) in a permanent-magnet synchronous machine (PMSM) degrades its energy efficiency and may cause other problems, accurate fault diagnosis for the PMSMs is necessary. In this paper, we present a method that uses voltage and current residual components to diagnose the phase and the severity of the ISF in PMSMs. A faulty-winding model (FWM) is introduced to analyze the fault; the model describes a relationship between the voltage decrease of the faulty winding to the characteristics of the fault such as a fault resistance and a fault turn ratio. Based on the model, we suggest a fault index (FI) that shows the severity of the ISF. As the ISF induces residual components to both voltage and current, the proposed method utilized those components with the FWM. A least squares (LS) method is used to estimate the FI and to diagnose the ISF of PMSMs. Experimental results demonstrate that the proposed method diagnose the ISF effectively.

Transmission line (TL) inspection is important for ensuring a stable supply of electricity to rural areas. Currently, there are several TL detection approaches based on computer vision; however, they have limitations owing to background clutter in visible light images. This article presents a novel multimodal image feature fusion module that utilizes both visible light and infrared images to enhance the TL-detection performance. The proposed module consists of a multibranch feature extraction (MFE) block followed by a channelwise attention (CA) block. The first block extracts the representative features of each modal input using multiple branches. The outputs of the MFE block are jointly aggregated into an attention vector in the CA block. Finally, the attention vector recalibrates each input feature of the proposed module. To reduce the number of additional parameters due to the insertion of the module, we introduced a channel-shrink factor in the MFE block and utilized a $1\times {1}$ convolution in the CA block. Comparison experiments with various augmented conditions of day, night, fog, and snow were conducted on a real-world dataset, which we constructed by visible light and infrared images. The results showed that the proposed module outperformed not only the case of single modal input but also the state-of-the-art fusion methods, regardless of the baseline networks. Additionally, the proposed module showed effectiveness in terms of capacity when the baseline network has a large number of weight parameters.

We present a method that detects and classifies demagnetization and interturn short faults (ISFs), which degrade the performance of permanent-magnet synchronous machines (PMSMs). Demagnetization and ISF are analyzed using models; we focused on changes of magnitude and angle of currents in the synchronous frame. Those two faults increase the magnitude of the input current compared with normal machines at the same load torque. However, our analysis suggests that demagnetization causes increase in the current angle β, whereas ISF causes decrease in β. We exploit this difference in response to classify the fault if one is detected in the PMSM. Experimental studies on interior-type PMSMs verify that demagnetization and ISF can be detected and classified in various operation conditions.