TY - JOUR
T1 - Effective landing strategy of robot leg using hybrid force/position control
AU - Kim, Kyung Hwan
AU - Kim, Jung Yup
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.
PY - 2022/11
Y1 - 2022/11
N2 - This paper deals with an effective strategy for the landing of a robot leg on uneven terrain. When the foot of the walking robot lands on uneven terrain, the actual ground contact point may be different from the reference point at every step. This may cause an early landing of the robot foot, resulting in a large landing impact force, or destabilize the robot’s balance as the foot fails to land at the planned time. To land on uneven terrain and generate a suitable vertical force to support robot weight, first, we propose a foot velocity and force transition control during the swing down and stance phases. The velocity control is used to land on the ground at various heights while reducing the landing impact, and the force control is used to support the robot’s weight after landing. The velocity control and the force control are transitioned to each other continuously based on the encoder feedback. Second, we propose a modified task-space computed torque control (CTC) to solve a coupling problem found in the conventional CTC during hybrid force/position control. Third, we propose a friction compensation method for implementing and verifying the landing strategy in real robots. The proposed friction compensation method uses the Coulomb-Viscous friction model, and the coefficients of the friction model are derived from the modified joint-space CTC with Linear Segments with Parabolic Blends (LSPB) trajectory. Finally, the landing strategy was verified through a 3-DOFs robot leg and generated the desired vertical force with a significantly small impact during landing.
AB - This paper deals with an effective strategy for the landing of a robot leg on uneven terrain. When the foot of the walking robot lands on uneven terrain, the actual ground contact point may be different from the reference point at every step. This may cause an early landing of the robot foot, resulting in a large landing impact force, or destabilize the robot’s balance as the foot fails to land at the planned time. To land on uneven terrain and generate a suitable vertical force to support robot weight, first, we propose a foot velocity and force transition control during the swing down and stance phases. The velocity control is used to land on the ground at various heights while reducing the landing impact, and the force control is used to support the robot’s weight after landing. The velocity control and the force control are transitioned to each other continuously based on the encoder feedback. Second, we propose a modified task-space computed torque control (CTC) to solve a coupling problem found in the conventional CTC during hybrid force/position control. Third, we propose a friction compensation method for implementing and verifying the landing strategy in real robots. The proposed friction compensation method uses the Coulomb-Viscous friction model, and the coefficients of the friction model are derived from the modified joint-space CTC with Linear Segments with Parabolic Blends (LSPB) trajectory. Finally, the landing strategy was verified through a 3-DOFs robot leg and generated the desired vertical force with a significantly small impact during landing.
KW - 3-DOF robot leg
KW - Computed torque control
KW - Friction compensation model
KW - Hybrid force/position control
KW - Landing strategy
UR - https://www.scopus.com/pages/publications/85137256762
U2 - 10.1007/s11370-022-00441-7
DO - 10.1007/s11370-022-00441-7
M3 - Article
AN - SCOPUS:85137256762
SN - 1861-2776
VL - 15
SP - 579
EP - 592
JO - Intelligent Service Robotics
JF - Intelligent Service Robotics
IS - 5
ER -