Manual Control of Bipedal Walking

After downloading bw_manual_app.jar, please execute it by double-clicking, or typing "java -jar bw_manual_app.jar".

To compare with "Reinforcement Learning for CPG-controlled Bipedal Walking",
let us try to control the identical model of bipedal walking manually.
You might understand its difficulty.

If the above application does not start, please install Java from

[How to control]
The model of biped walking (Taga, 1991) is written by an equation of motion of 14 variables
with 8 constraints.
The position of the model is indicated by 6 variables because 14 - 8 = 6, and they are the position (x, y) of the hip, and the angles of hips and knees (θR1, θR2, θL1, θL2) for both feet.

Moreover, by considering their time-derivatives (vx, vy) and (ωR1, ωR2, ωL1, ωL2), the state of this dynamical system can be described.
The torque T given to this model is determined by PD control scheme.
In this applet, the destination angles for both feet θR1d and θL1d can be changed with two horizontal scroll bars.

θR2d and θL2d are determined automatically so that θR2d=55 when θR1d > 0 and ωR1d > 0 are satisfied, and otherwise θR2d=0.

This page is based on the following paper.

<< Reinforcement Learning for CPG-controlled Bipedal Walking / Cart-Pole Balancing by Reinforcement Learning >>

Back to Takashi Kanamaru's Web Page