Humanoid Robot, RoK-3
RoK-3
RoK-3 is Humanoid Robot of Kookmin University. The height of the robot standing with two legs fully stretched out is about 127 cm, its weight is about 44 kg, and the total degree of freedom (DoFs) is 13. The joint controller of the robot works based on current control or position control.
The robot is used primarily to study the field of biped walking. The robot can walk at speeds of about 1 km/h or more. Using the robot, we also study controls such as balancing control to achieve robust and stable walking on uneven terrain. Furthermore, to implement robust walking, we study push recovery that can deal with disturbances applied to the robot.
RoK-3 walking on the treadmill
RoK-3 Balancing Control on Slope
RoK-3
Specifications
Standing Height : 127cm
Weight : 44kg
Using 17 Brushless DC Motors
Total 13 DoF
Research achievements with RoK-3
Humanoid Robot, RoK-3
RoK-3 can be a full body humanoid robot by combining RoK-3 in the form of biped robots with the RoK-Arm9 which is the upper body of RoK-3.
RoK-3 (Humanoid Robot)
RoK-3 (Biped Robots)
RoK-Arm9
RoK-3 walking on the cinder blocks with footstep algorithm
RoK-3 can stably cross cinder blocks with an area acquired from a camera. The robot uses the point cloud obtained from the camera to detect the planes and calculate their centroids and directions. This information is used to determine the position and orientation of the foot to which the robot should go. It updates the local map information with every walking step to quickly recognize nearby objects and find a path that can move.
Walking on the cinder blocks with footstep algorithm
RoK-3, Slope Observer & Control based Walking on Unknown Slope
We develop the slope observer which can estimate the inclined angle and angular velocity of the COM and the slope angle of terrain to realize stable biped walking on a slope without the IMU sensor. We develop the proposed slope observer-based posture-stabilization algorithm. The proposed observer and posture-stabilization algorithm were verified through a slope walking experiment using the real humanoid robot RoK-3.
Therefore, our slope-walking algorithm based on the proposed observer and posture stabilization can be used as an alternative when the IMU sensor-based slope-walking algorithm does not work due to an unexpected problem with the sensor.
Slope Observer & Control based Walking on Unknown Slope
RoK-3 walking on uneven terrain with Variable Stiffness Sole
The present study developed the Variable Stiffness Sole (VSS) module, which can adapt to the irregularity of the ground and provide robust support by changing the stiffness. The developed VSS module can be applied to a real humanoid robot. VSS consists of a silicone vessel with a flat hexahedral shape and an ankle mount for jointing with the ankle link of the robot.
Next, we developed a walking control framework using VSS. The proposed walking control framework comprises a walking pattern generator, controllers for walking stabilization, and a VSS controller for adapting to the ground shape and supporting the robot robustly when the foot lands on the ground.
The VSS controller changes the stiffness of the VSS based on force and torque (FT) sensors at the ankle and a walking pattern scheme. The effectiveness of the proposed walking control framework using VSS be verified through the obstacle stepping experiment and a walking experiment on uneven terrain. The experiment is conducted with our real humanoid robot RoK-3 equipped with the VSS module.
RoK-3 walking on uneven terrain with VSS (Variable Stiffness Sole)
RoK-3 equipped with VSS