Torso servoing significantly reduces the magnitude of the excursions in the forward dimension

and, as a side effect, can also reduce the excursions in the lateral dimension.

Figure 4.15 shows

the results for a set of trials which use the same balance control parameters as the walks of Figure

4.3.

In this set of trials,

torso servoing is applied using a desired angle of 5 degrees forward

from vertical.

The resulting paths are straighter and generally more consistent in direction than the non-torso

servoed results. The useful range of Q^dincreases in general.

The phase diagram in Figure 4.16

illustrates the effectiveness of torso servoing at reducing the

bobbing

effect

caused

by

the

particular stance hip perturbation used.

In this trial, torso servoing reduces the range of torso

pitch from approximately 12 degrees to 3 degrees.

Q_lat

Figure 4.16-

Continuous-time up vector component phase diagram for
human model with torso servoing to +5 degrees from
world vertical (Q_servo[!] .09). Compare to Figure 4.2.

4. 5

Robo-bird Running

Figure 4.17 shows the base PCG used to generate a running motion for the robo-bird model,

shown in Figure 4.18.

While the base PCG differs from that used for the human model, it is

balanced using up-vector RVs and stance hip pitch and roll perturbations as with the human

model.