The image
data are sent to an image grabber, where image processing is performed
to determine the shape and size of any obstacles. If an
obstacle is found 2, the robot must decide whether the robot to
detour around it or jump over it based on its geometrical measurements,
which are obtained by integrating the camera images and
sensor information. When a decision about the object is made,
the final command is delivered to the robot controller digital signal
processor. If the obstacle is small enough to be overcome by jumping,
the robot aligns itself so that its direction of motion is normal
to the obstacle baseline. After estimating the height of the obstacle,
the jumping velocity, which determines the jumping height, is calculated
and a duty ratio for the valve that controls the pneumatic
cylinder is determined. Eventually, the robot jumps up and lands
on top of the obstacle, or jumps over the obstacle if it is sufficiently
low 3. If the object is too high for robot to jump up on it, the robot
detours around the obstacle and keeps moving afterward 4.