Learning of Motion Primitives Using Reference-Point-Dependent GP-HSMM for Domestic Service Robots
In this paper, we propose a method for motion learning aimed at the execution of autonomous household chores by home robots in real environments. For robots to act autonomously in a real environment, it is necessary to define appropriate actions for the environment. However, it is difficult to define these actions manually. Therefore, body motions that are common to multiple actions are defined as motion primitives. Complex actions can then be learned by combining these motion primitives. For learning motion primitives, we propose reference-point-dependent Gaussian process hidden semi-Markov model (RPD-GP-HSMM). For verification, a robot is tele-operated in order to perform actions included in several domestic household chores. The robot then learned the associated motion primitives from the robot’s body information and object information.
KeywordsMotion learning Gaussian process Hidden semi-Markov model Reference-point
- 3.Sanzari, M., et al.: Human motion primitive discovery and recognition. arXiv preprint arXiv:1709.10494 (2017)
- 4.Kuniyoshi, Y., et al.: Learning by watching: extracting reusable task knowledge from visual observation of humanGoogle Scholar
- 6.Nakamura, T., Nagai, T., Mochihashi, D., Kobayashi, I., Asoh, H., Kaneko, M.: Segmenting continuous motions with hidden semi-Markov models and Gaussian processes. Frontiers Neurorob. 11 (2017)Google Scholar
- 7.Uchiumi, K., Tsukahara, H., Mochihashi, D.: Inducing word and part-of-speech with pitman-yor hidden semi-Markov models. ACL 1, 1774–1782 (2015)Google Scholar