Explore chapters and articles related to this topic
Towards physical interaction-based sequential mobility assistance using latent generative model of movement state
Published in Advanced Robotics, 2021
Shunki Itadera, Taisuke Kobayashi, Jun Nakanishi, Tadayoshi Aoyama, Yasuhisa Hasegawa
In this paper, we have introduced our development of assistive mobility aid for supporting user's movement in their daily life such as movement transition and walking. First, we proposed an unsupervised learning model based online estimator of the user's state using an LSTM based VAE in order to recognize user's movement state in a lower-dimensional latent space. Then, we presented an impedance control-based assistive strategy on a mobile assistive robot in order to provide a user with an appropriate amount of physical support according to the current user's movement and transition state. Finally, we demonstrated the effectiveness of our approach in the online classification of user's movement states and experimental results in providing assistive mobility aid using Toyota Human Support Robot including novel movement detection.