A very interesting paper on
In-flight measurement of kite deformations with inertial sensors
Master’s thesis report by Swami Jonard
Crikey that’s some thorough work
1 Like
On my comment, I quote a figure (here it is figure 7.6). As two methods of measurement are mentioned (“photogrammetric measurement” and “inertial measurement”) I tried to connect said figure with texts about both to help to understand.
Figure 7.6, among
List of Figures
7.6 Example of the misalignment between the orientations of the two kite shape
measurements at the peak of a right turn during a looping sequence (166.8 s).
The photogrammetric measurement is shown in blue and the inertial measurement is shown in orange. The photogrammetric shape estimate appears
to be pitched down compared to the inertial estimate, despite having the
same shape. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Page 9:
2.2.2. Inertial measurement
Inertial measurement units (IMU) that incorporate a gyroscope, an accelerometer, and sometimes a magnetometer are also used to estimate a wing’s shape. These sensors can be
lightweight and easier to set up than strain sensors and can directly provide information
on the orientation and position of a point on the wing. The position and orientation are
determined in a way analogous to inertial navigation systems (INS), by using sensor fusion
algorithms such as Kalman filters to compensate for the integration drift of acceleration
and angular velocity (further explored in section 2.3). (Gong et al., 2017) These algorithms
are computationally light and with the development of micro-electro-mechanical systems
(MEMS), inertial sensors have become very compact and cheap while delivering a decent performance. Inertial sensing is therefore more suited for real-time shape estimation (Schirmer
& Steck, 2020; Woodman, 2007).
Page 45: chapter 5 about photogrammetry:
Photogrammetry
As seen in the literature study, photogrammetry is the main established technique used for
kite deformation measurement. However, its applications are limited to close-range controlled experiments. For outdoor applications, the constraints of practicality impose a setup
where a minimal number of cameras are used and the resolution of the cameras limit the
distance at which the object is measured. In this study, photogrammetry serves the purpose
of calibrating the sensor alignment and validating inertial sensor data. Point tracking will
be used for the 2-D tracking since it is simple to implement, using the open-source software
Blender as well as some post processing. This method provides a position measurement of
discrete points which can be compared directly to the sensor positions estimated by the
algorithm presented in chapter 4. In this chapter, the different steps to obtain photogrammetric position measurements are presented.
For me the research reassuringly highlighted how useful IMU sensors can be in kite attitude data.
One disappointment for me was how little time was dedicated to field testing prototypes.
A simple problem with a duct taped sensor coming away from the struts … significantly affected the reults.
There was another article published recently which may (in the future caveat) help alleviate the stick on problem… A soft CPU circuit
https://www.nature.com/articles/s41586-021-03625-w
While the tech is still limited at this stage…
That sort of compute tech could be bonded more effectively to kite skins