This paper proposes a deep learning-based anomaly detection method using time-series vibration and current data, which were obtained from endurance tests on driving modules applied in industrial robots and machine systems. Unlike traditional classification models that depend on labeled fault data for detection, acquiring sufficient fault data in real industrial environments is highly challenging due to various conditions and constraints. To address this issue, we employ a semi-supervised learning approach that relies solely on normal data to effectively detect abnormal patterns, overcoming the limitations of conventional methods. The performance of semi-supervised models was first validated using a statistical feature-based anomaly detection approach, from which the GCN-VAE model was adopted. By combining the spatial feature extraction capability of Graph Convolutional Networks (GCNs) with the latent temporal feature modeling of Variational Autoencoders (VAEs), our method can effectively detect abnormal signs in the data, particularly in the lead-up to system failures. The experimental results confirmed that the proposed GCN-VAE model outperformed existing hybrid deep learning models in terms of anomaly detection performance in the pre-failure section.

This paper proposes a safety response technology that leverages multi-domain vision sensors to monitor the positions of workers and robots in real-time. The technology calculates flexible protective separation distances (FPSD), which facilitate dynamic control of robot velocity and enable safety stops. Compliant with the ISO/TS 15066 technical specifications for speed and separation monitoring, this method not only satisfies stringent safety standards but also promotes a more adaptable working environment and improves productivity. The proposed method has been analyzed and validated using data acquired from multi-domain sensors within the collaborative workspace of worker-robot collaboration tasks, and its operation has been confirmed at variable worker and robot velocities through experimental scenarios.

This paper proposes fault diagnosis methods aimed at proactively preventing potential safety issues in robot systems, particularly human coexistence robots (HCRs) used in industrial environments. The data were collected from durability tests of the driving module for HCRs, gathering time-series vibration data until the module failed. In this study, to apply classification methods in the absence of post-failure data, the initial 50% of the collected data were designated as the normal section, and the data from the 10 h immediately preceding the failure were selected as the fault section. To generate additional data for the limited fault dataset, the Wasserstein generative adversarial networks with gradient penalty (WGAN-GP) model was utilized and residual connections were added to the generator to maintain the basic structure while preventing the loss of key features of the data. Considering that the performance of image encoding techniques varies depending on the dataset type, this study applied and compared five image encoding methods and four CNN models to facilitate the selection of the most suitable algorithm. The time-series data were converted into image data using image encoding techniques including recurrence plot, Gramian angular field, Markov transition field, spectrogram, and scalogram. These images were then applied to CNN models, including VGGNet, GoogleNet, ResNet, and DenseNet, to calculate the accuracy of fault diagnosis and compare the performance of each model. The experimental results demonstrated significant improvements in diagnostic accuracy when employing the WGAN-GP model to generate fault data, and among the image encoding techniques and convolutional neural network models, spectrogram and DenseNet exhibited superior performance, respectively.

In the wake of COVID-19, the digital fitness market combining health equipment and ICT technologies is experiencing unexpected high growth. A smart trampoline fitness system is a new representative home exercise equipment for muscle strengthening and rehabilitation exercises. Recognizing the motions of the user and evaluating user activity is critical for implementing its self-guided exercising system. This study aimed to estimate the three-dimensional positions of the user’s foot using deep learning-based image processing algorithms for footprint shadow images acquired from the system. The proposed system comprises a jumping fitness trampoline; an upward-looking camera with a wide-angle and fish-eye lens; and an embedded board to process deep learning algorithms. Compared with our previous approach, which suffered from a geometric calibration process, a camera calibration method for highly distorted images, and algorithmic sensitivity to environmental changes such as illumination conditions, the proposed deep learning algorithm utilizes end-to-end learning without calibration. The network is configured with a modified Fast-RCNN based on ResNet-50, where the region proposal network is modified to process location regression different from box regression. To verify the effectiveness and accuracy of the proposed algorithm, a series of experiments are performed using a prototype system with a robotic manipulator to handle a foot mockup. The three root mean square errors corresponding to X, Y, and Z directions were revealed to be 8.32, 15.14, and 4.05 mm, respectively. Thus, the system can be utilized for motion recognition and performance evaluation of jumping exercises.

Classifying low-concentration Gas Sensor Array (GSA) data is hard due to low SNR, sensor heterogeneity, drift, and small samples. We benchmark time-series-to-image (TS2I) CNNs against time-series classifiers, after reproducing a strong FE-ELM baseline under a shared fold manifest. Using the GSA-LC and GSA-FM datasets, we compare FE-ELM, vector baselines, time-series methods, and TS2I-CNNs with 20 × 5 repeated stratified cross-validation (n = 100). ROCKET delivers the best accuracy on both datasets and is significantly better than TCN and MiniROCKET (paired tests with Holm–Bonferroni, p < 0.05): on GSA-FM, accuracy 0.9721 ± 0.0480 (95% CI [0.9627, 0.9815]) with Macro-F1 0.9757; on GSA-LC, 0.9578 ± 0.0433 (95% CI [0.9493, 0.9663]) with Macro-F1 0.9555. Among image-based models, CNN-RP is the most robust, whereas CNN-GASF lags, especially on GSA-LC. RGB fusion strategies (e.g., with MTF) are dataset-dependent and often inconsistent, and transfer learning with ResNet-18 offers no consistent advantage. Overall, ROCKET ranks first across folds, while CNN-RP is the most reliable TS2I alternative under low-concentration conditions. These results provide a reproducible, fair benchmark for e-nose applications and practical guidance for model selection, while clarifying both the potential and limitations of TS2I.

A software robot requires plausible external features for intimate interaction with humans. The various appearance designs effectively contribute to provide him/her with an opportunity to select a preferable robot among them. This paper proposes genetic representation for the appearance of software robot, which is inspired by homologous chromosomes. User selection scheme is based on the interactive genetic algorithm. As user evaluates the preference of external features of software robot, proposed algorithm converges to the most preferable appearance to user's choice. The effectiveness of the proposed scheme is demonstrated through experiments.