Fig 3 - uploaded by Visak Kumar
Content may be subject to copyright.
Download
View publication
Illustration of locomotion policies deployed on the Darwin OP2 robot. Top: walk forward. Middle: walk backward. Bottom: walk sideways.

Illustration of locomotion policies deployed on the Darwin OP2 robot. Top: walk forward. Middle: walk backward. Bottom: walk sideways.

Source publication
Fig. 1. The simulated and the real Darwin OP2 robot trained to walk...
Fig. 2. Overview of the proposed algorithm.
Fig. 3. Illustration of locomotion policies deployed on the Darwin OP2...
Fig. 4. Comparison of the distance travelled by the robot using our...
+2 Fig. 5. Comparison of the elapsed time before the robot loses balance...
Sim-to-Real Transfer for Biped Locomotion
Preprint
Full-text available
  • Mar 2019
We present a new approach for transfer of dynamic robot control policies such as biped locomotion from simulation to real hardware. Key to our approach is to perform system identification of the model parameters {\mu} of the hardware (e.g. friction, center-of-mass) in two distinct stages, before policy learning (pre-sysID) and after policy learning...
Cite
Download full-text

Contexts in source publication

Context 1
... robot is tasked to walk on a yoga mat that lies on top of a white board, as shown in Figure 3. We choose this deformable surface to better provide protection for the robot. ...
Context 2
... results in a faster walking gait, and may also have contributed to the larger variance seen in our policy. An illustration of the three locomotion tasks with our trained policies can be seen in Figure 3. ...
Context 3
... robot is tasked to walk on a yoga mat that lies on top of a white board, as shown in Figure 3. We choose this deformable surface to better provide protection for the robot. ...
Context 4
... results in a faster walking gait, and may also have contributed to the larger variance seen in our policy. An illustration of the three locomotion tasks with our trained policies can be seen in Figure 3. ...

Similar publications

Figure 1 -Simple schematic of robotic test facility
Figure 4 -Step Responses For Conventional Controller, Varying K e .
Figure 5 -Fuzzy Control Using SFAC
Figure 6 -Step Response For Fuzzy Force Controller, Varying K e .
Validation of several novel force controller techniques with an open architecture Robot controller
Conference Paper
Full-text available
  • Apr 2001
This paper describes a research programme currently underway at the University of Sunderland developing a new, PC-based, open architecture robot controller for a new generation of robots. The hardware/software architecture under development is presented. This paper also introduces a range of novel techniques for the design of intelligent force cont...
View