A twin-armed robot by Professor Tetsuya Ogata and his colleagues at Waseda University
Professor Tetsuya Ogata and his teamat Waseda University have developed a technology that allows a deep learning model to learn and perform multiple actions. They demonstrated the function by applying it to a twin-armed robot folding towels.In the case of folding towels, four motions were learned, including folding the towel from the edge and repositioning the towel. We trained the deep learning model so that the internal state of the model is the same at the beginning and end of each action. For each of the four actions, we used a large deep learning model with about 300 cells and 400 steps. In this study, we learned multiple actions by connecting them to a single model. In the future, we will apply the model to complex operations such as skewering to verify the learning method suitable for each operation. The technology for using deep learning to identify images is already in the practical stage. Photo taken on October 16, 2019, credit: Nikkan Kogyo Shimbun / Kyodo News Images
- Product Code
- ILEA000687953
- Registered date
- 2019/10/16 00:00:00
- Credit
- THE NIKKAN KOGYO SHIMBUN / Kyodo News Images
- Media source
- THE NIKKAN KOGYO SHIMBUN
- Media size
- 1944 × 2592 pixel
- Resolution
- 350 dpi
- Deployment size
- 2.21(MB)*
- Special instruction
- **The text may be generated by an automatic translation system**
*File size when opened in Photoshop, etc.
