In collaboration with Frontec, which produces parts such as bolts and nuts for the automobile industry, Kyung Hee University and Benple Inc. develop and deploy AI system for automatic quality inspection of weld nuts. Various constraints to consider exist in adopting AI for the factory, such as response time and limited computing resources available. Our convolutional neural network (CNN) system using large-scale images must classify weld nuts within 0.2 seconds with accuracy over 95%. We designed Circular Hough Transform based preprocessing and an adjusted VGG (Visual Geometry Group) model. The system showed accuracy over 99% and response time of about 0.14 sec. We use TCP / IP protocol to communicate the embedded classification system with an existing vision inspector using LabVIEW. We suggest ways to develop and embed a deep learning framework in an existing manufacturing environment without a hardware change.

For the last few years, there has been a large volume of research on Peer-to-Peer (P2P) system, resulting in many hybrid P2P models. This model is important to keep a high available SP because all P2P overlay traffic will be routed via the SP. Main purpose of this paper is to improve SP's availability for OPs so that the system selects the peer with the potential to go the longest without needing replacement. In order to achieve the goal, we study a peer's availability prediction based on analysis of SP's characteristics and past behavior. Based on these, we propose various replacement policies and evaluate them through experiments. This paper can improve the resilience of the system by improving the availability of SPs selected by OPs. Second, we analyze inherent SP's characteristics and its past behavior and discuss which factors affect the performance of the system. Finally, we show that predicted residual lifetime can help us understand the benefits of replacement policies.