Giovanni De Magistris

Spatio-Temporal Anomaly Detection for Industrial Robots through Prediction in Unsupervised Feature Space

Spatio-temporal anomaly detection by unsupervised learning have applications in a wide range of practical settings. In this paper we present a surveillance system for industrial robots using a monocular camera. We propose a new unsupervised learning method to train a deep feature extractor from unlabeled images. Without any data augmentation, the algorithm co-learns the network parameters on different pseudo-classes simultaneously to create unbiased feature representation. Combining the learned features with a prediction system, we can detect irregularities in high dimensional data feed (e.g. video of a robot performing pick and place task). The results show how the proposed approach can detect previously unseen anomalies in the robot surveillance video. Although the technique is not designed for classification, we show the use of the learned features in a more traditional classification application for CIFAR-10 dataset.

Human-Like Hand Reaching by Motion Prediction Using Long Short-Term Memory

An interaction between a robot and a human could be difficult with only reactive mechanisms, especially in a social interaction, because the robot usually needs time to plan its movement. This paper discusses a motion generation system for humanoid robots to perform interactions with human motion prediction. To learn a human motion, a Long Short-Term Memory is trained using a public dataset. The effectiveness of the proposed technique is demonstrated by performing a handshake with a humanoid robot. Instead of following the human palm, the robot learns to predict the hand-meeting point. By using three metrics namely the smoothness, timeliness, and efficiency of the robot movements, the experimental results of various motion plans are compared. The predictive method shows a balanced trade-off point in all the metrics.