 |
| Titel |
Smartstones: a small e-compass, accelerometer and gyroscope embedded in stones |
| VerfasserIn |
Oliver Gronz, Priska H. Hiller, Stefan Wirtz, Kerstin Becker, Thomas Iserloh, Jochen Aberle, Markus C. Casper |
| Konferenz |
EGU General Assembly 2015
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 17 (2015) |
| Datensatznummer |
250112057
|
| Publikation (Nr.) |
 EGU/EGU2015-12214.pdf |
|
|
|
|
|
| Zusammenfassung |
| Pebbles or rock fragments influence soil erosion processes in various ways: they can protect
the soil but also enhance the erosion as soon as they are moved by water and impact onto soil.
So far, stone-embedded devices to measure the movements have been quite big, up to several
decimetres, which does not allow for the analysis of pebbles from medium and coarse gravel
classes. In this study, we used a novel device called Smartstones, which is significantly
smaller.
The Smartstone device’s dimensions are 55 mm in length, 8 mm in diameter and an
approximately 70 mm long flexible antenna (device developer: SMART-RFID solutions
Rheinberg, Germany). It is powered by two button cells, contains an own data storage and is
able to wait inactive for longer times until it is activated by movement. It communicates
via active RFID (radio frequency identification) technology to a Linux gateway,
which stores the sensor data in a database after transmission and is able to handle
several devices simultaneously. The device contains a Bosch sensor that measures
magnetic flux density, acceleration and rotation, in each case for / around three
axes.
In our study, the device has been used in a laboratory flume (270 cm in length, 5°
to 10° slope, approx. 2 cm water level, mean flow velocities between 0.66 and 1
ms-1) in combination with a high speed camera to capture the movement of the
pebbles. The simultaneous usage of two capture devices allows for a comparison
of the results: movement patterns derived from image analysis and sensor data
analysis.
In the device’s first software version, all three sensors – acceleration, compass, and
gyroscope – were active. The acquisition of all values resulted in a sampling rate of 10 Hz.
After the experiments using this setup, the data analysis of the high speed images and the
device’s data showed that the pebble reached rotation velocities beyond 5 rotations per
second, even on the relatively short flume and low water levels. Thus, the device produced
only sub-Nyquist sampling values and the rotation velocity of the pebble could not be derived
correctly using solely the device’s data.
Consequently, the device’s software was adapted by the developers: the second (and
current) version of the device only acquires acceleration and compass, as the acquisition of
the gyroscope’s value does not allow for higher sampling rates. The second version samples
every 12 ms. All aforementioned experiments have been repeated using the adapted device.
For data analysis, the high-speed camera images were merged with the device data using a
MATLAB script. Furthermore, the derived relative pebble orientation – yaw, pitch and roll –
is illustrated using a rotated CAD model of the pebble. The pebble’s orientation is derived
from compass and accelerometer data using sensor fusion and algorithms for tilt compensated
compasses.
The results show that the device is perfectly able to capture the movement of the pebble
such as rotation (including the rotation axis), sliding or saltation. The interacting forces
between the pebble and the underground can be calculated from the acceleration data.
However, the accelerometer data also showed that the range of the sensor is not sufficiently
large: clipping of values occurred. According to present instrument specifications, the sensor
is able to capture up to 4 g for each axis but the resulting vectors for acceleration along all
three axes showed values greater than 4 g, even up to the theoretical maximum of
approximately 6.9 g. Thus, an impact of this strength that only stresses one axis cannot
be measured. As a result of this clipping, the derivation of the pebble’s absolute
position using double integration of acceleration values is associated with flaws.
Besides this clipping, the derived position will deviate from the true position for larger
distances or longer experiment durations as the noise of the data will be integrated,
too.
Several requirements for the next device version were formulated:
The range of the accelerometer will be set to the sensor’s maximum of 16 g.
The device will be water proof.
Data analysis will include further methods like Hidden Markov Models or
Kalman Filtering as the tilt-compensation is actually not intended for irregular
moving devices. These techniques are well-established for other devices and
purposes like navigation using GPS.
In near future, the Smartstone device will be used outside the laboratory in natural rills
and rill experiments. In these experiments, the water is turbid and the pebble will not be
visible at all, which does not allow for the usage of the high speed camera. However, the
present results showed that the movement of the pebble in addition to the applied forces to
the underground and the rill’s sidewalls can be captured solely by the Smartstone. |
|
|
|
|
|
|