Robot Cooperation on Peer Reciprocal Learning and IoT-based Automatic Packaging

• Main Authors: Chih-YinLiu, 劉致吟
• Other Authors: Tzuu-Hseng S. Li
• Format: Others
• Language: en_US
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/9nnp4s

博士 === 國立成功大學 === 電機工程學系 === 105 === In this dissertation, we propose two systems of robot cooperation, a robot peer reciprocal learning system and an Internet-of-Things (IoT)-based automatic e-fulfillment packaging system, based on two different perspectives. The robot peer reciprocal learning system is inspired by human peer learning theories. In this system, similar function robots are treated as learning partners, they can not only cooperatively accomplish difficult tasks but also help each other to learn better. For learning a new concept, we also propose a mutual concept learning method which allows the robots to learn from each other by exchanging weights in their neural network concept learning system. On the other hand, the IoT-based automatic e-fulfillment packaging system is an implementation of the concept of Industry 4.0. In the system, all components are connected and monitored by a cyber network, and cooperatively accomplish a packaging task in a warehouse e-fulfillment situation. For packing all items inside a suitably sized box, we propose a 3D Adaptive Particle Swarm Optimization (PSO)-based packing algorithm. Finally, the simulations and the experiments demonstrate the efficiency of the proposed robot cooperation systems and learning algorithms.