Title
Multi-Contacts Force-Reactive Walking Control during Physical Human-Humanoid Interaction
Abstract
This paper proposes a force-reactive walking control framework for humanoid robots, which consists of ankle and stepping strategies for stabilization during physical human-robot interaction (pHRI). Contact forces measured by robotic skin are explicitly considered in the proposed framework. The ankle strategy aims to keep the footstep location and duration within a nominal range, and this strategy is used to generate physical cues to inform the human partner about the robot's walking intentions. Thereafter, only when the ankle strategy cannot keep the walking balance, the stepping strategy adjusts the footstep location, the duration of single support phase, and the timing to start the next step to follow the partner's intention, which is derived via the contact interaction. By exploiting these two strategies, pHRI can be achieved where intentions of the partner and the robot are mutually exchanged. In a real experiment with a full-sized humanoid robot, where the partner teaches the footstep to the robot, our proposed framework succeeded in adjusting the footstep according to multiple contact forces related to the partner's intention, while basically trying to keep its own nominal footstep.
Year
DOI
Venue
2019
10.1109/Humanoids43949.2019.9035078
2019 IEEE-RAS 19th International Conference on Humanoid Robots (Humanoids)
Keywords
DocType
ISSN
multicontacts force-reactive walking control,physical human-humanoid interaction,force-reactive walking control framework,humanoid robot,stepping strategies,human-robot interaction,pHRI,robotic skin,ankle strategy,footstep location,nominal range,human partner,walking intentions,walking balance,stepping strategy,single support phase,contact interaction
Conference
2164-0572
ISBN
Citations 
PageRank 
978-1-5386-7631-8
1
0.35
References 
Authors
0
5
Name
Order
Citations
PageRank
Taisuke Kobayashi13110.92
Emmanuel C. Dean-Leon26215.39
Julio Rogelio Guadarrama-Olvera310.35
Florian Bergner4216.14
Gordon Cheng51250115.33