allows the robot to react to uneven terrain can be
challenging and time consuming all by itself. It can also
be very rewarding. The feeling of success you get when the
simulated robot takes a few steps on its own can be just as
satisfying as programming the real thing.
If you would like to see the simulation in action before
you try any programming, you can search You Tube for
“RobotBASIC Walking Humanoid” and watch a short video
showing the concepts discussed here. The video concludes
with the demonstration of a simple algorithm that allows
the robot to autonomously handle a wide variety of
The algorithm is based on the idea that an autonomous
robot has to repeat left and right steps as shown by the
code in Figure 8. In order to perform properly, the LtStep
and RtStep subroutines must keep the robot balanced while
constantly adjusting the leg angles to handle
unknown slopes. Additionally, each of these routines
must terminate with the robot returning to a
predetermined normalized position so that the next
step starts from a known condition. Without this
normalized position, even small errors will add up over
time and cause an imbalance that cannot be corrected.
The StartWalk routine should move the robot to the
normalized position while the Stop Walk must find a
balanced standing position.
A realistic simulation can let you experiment with
autonomous walking algorithms without destroying
your robot. SV
Forward The robot is off balance and falling forward.
Backward The robot is off balance and falling backward.
Vertical The robot is near vertical (and balanced).
Balanced The robot is forward but still balanced.
Figure 8. This code makes the robot walk across
the screen, but only if you have the proper subroutines.
Complete Fabrication Center
AS 9100 Registration
SERVO 11.2009 41