Biomedical Engineering Reference
In-Depth Information
FIGURE 9.24
If navigation is done purely by heading, wind causes the direction of travel to be the resultant of the true
airspeed along the heading direction and the wind vector.
Wind may not be a significant issue for a fast
platform; however, the forward velocity of a
micro UAV might approach 0. A simple feed-
back loop attempting to reduce altitude to
increase optical flow to some expected value
might reduce altitude until contact with the
ground occurs.
The difficulties of computing height from a
single optical-flow sensor at an instant are
intractable. Obvious solutions, such as
modulating speed and observing changes in
direction of flow or magnitude of flow, will
succeed under some conditions but fail under
others. In the experiments described in the
remainder of this section, when altitude was
being computed,
|
v
|
was the speed output of a
GPS. A fast or robust system could ignore the
GPS velocity and rely entirely on
v
TAS
with a
potentially insignificant reduction in the
accuracy of altitude flown.
provide the direction of flight. The direction of
flight is calculated from the sum of the magnetic
heading vector angle and the optical-flow vector
angle; thus,
∠
v
=
∠
m
−
∠
f
.
(9.18)
Obviously, the solution to following a desired
path is to fly the desired heading on the bearing
reference
∠
m
, measure
∠
f
, and compensate the
desired heading. This could be considered as a
single step, repeated at discrete intervals of min-
utes or seconds, as is appropriate compared to a
human pilot. The alternative is to run the system
as a closed-loop regulator, with the difference
between the actual heading angle and the desired
heading direction as the error to be nulled. The
regulator shown in
Figure 9.25
was implemented
in the flight test. A simple proportional/integral/
differential (PID) regulator was used to manage
the bank angle that causes change of heading.
9.5.3 Course
Correction of the course over the ground is pos-
sible using optical flow and a heading reference.
Substituting from Eq.
(9.14)
, we get
9.5.4 Flight Test
The small UAV used in the experiment is shown
in
Figure 9.26
. Optical flow was implemented
using a single sensor from an optical computer
mouse. The sensor provides a reliable optical-
flow signal under most daylight conditions.
It was designed for high optical-flow rates (as
would be experienced within 1 mm of a desk-
top). The optics we fitted had a narrow 5° field
m
v
TAS
|
m
|
+
w
kr
f
=−
(9.17)
;
hence, any changes in
r
have no effect on the
angle
∠
f
, but only on the magnitude
|
f
|
. Thus
measurements of optical-flow direction and
magnetic compass (or gyro compass) heading
