This

Rear view

Right side view

DOF:

Value:

1

0

2

0

3:0

0

3:1

0

4

0

5:0

0

5:1

0

6:0

-5.1

6:1

-20

7

50

8:0

0

8:1

0

Choice of Regulation Variables, Q

3. 5

As discussed earlier, the discrete balance control of Eq. 3.6 is applied only once per step.

implies that each regulation variable must represent the behaviour of some part of the system state

over an entire step as a single scalar value. It is important to choose suitable functions of state and

sampling times which give a reasonably smooth response to perturbations as we assume in

Section 3.2 by using the discrete system Jacobian.

All of the RVs presented in this thesis are projections of the system state at a specific point in the

cycle as shown in Figure 3.11.

The sample times correspond to the approximate time of foot

placement for each step. Variations such as sampling an average or peak value of some function

of state over the whole step are also promising but unexplored possibilities.