Recently, the estimation of joint kinetics such as joint force and moment using wearable inertial sensors has received great attention in biomechanics. Generally, the joint force and moment are calculated though inverse dynamics using segment kinematic data, ground reaction force, and moment. However, this approach has problems such as estimation error of kinematic data and soft tissue artifacts, which can lead to inaccuracy of joint forces and moments in inverse dynamics. This study aimed to apply a recurrent neural network (RNN) instead of inverse dynamics to joint force and moment estimation. The proposed RNN could receive signals from inertial sensors and force plate as input vector and output lower extremity joints forces and moments. As the proposed method does not depend on inverse dynamics, it is independent of the inaccuracy problem of the conventional method. Experimental results showed that the estimation performance of hip joint moment of the proposed RNN was improved by 66.4% compared to that of the inverse dynamics-based method.

The automated inspection method of paper cups by using a deep learning classifier is proposed. Unlike conventional inspection methods requiring defect detection, feature extraction, and classification stages, the proposed method gives a unified inspection approach where three separate stages are replaced by one deep-learning model. The images of paper cups are grabbed using a CCD (Charge Coupled Device) camera and diffused LED lights. The defect patches are extracted from the gathered images and then augmented to be trained by the deep- learning classifier. The random rotation, width and height shift, horizontal and vertical flip, shearing, and zooming are used as data augmentation. Negative patches are randomly extracted and augmented from gathered images. The VGG (Visual Geometry Group)-like classifier is used as our deep-learning classifier and has five convolutional layers and max-pooling layers for every two convolutional layers. The drop-outs are adopted to prevent overfitting. In the paper, we have tested four kinds of defects and nondefects. The optimal classifier model was obtained from train and validation data and the model shows 96.5% accuracy for test data. The results conclude that the proposed method is an effective and promising approach for paper cup inspection.