CFR-19-PerpetuumMobile

Re-landing a CanSat with a drone landing system

Main Author’s Name

Marius Glaeser (Elisabeth-Gymnasium), Germany,

Co- Author’s Names

Eric Oliver Schmidt (Georg-Cantor-Gymnasium), Germany,
Oliver Zimmermann (Georg-Cantor-Gymnasium), Germany,
Malte Lippert (Elisabeth-Gymnasium), Germany,
Fabian Schmidtchen (Elisabeth-Gymnasium), Germany,
Cedric Frauendorf (Elisabeth-Gymnasium), Germany,

Abstract
The following paper is a complete report from the CanSat 2019 Launch Campaign, which occurred between
24th and 28th of June 2019 in Castel San Pietro Terme, Bologna, Italy.
A CanSat is a "satellite" the size of a beverage can, developed, built and programmed by student teams. This
mini-satellite is brought to a height of several hundred meters by a rocket and then sinks to the ground on a
parachute. The satellite will perform measurements and other tasks for which it was designed.
We are carrying out a technical demonstration of a special landing system and build a CanSat that lands gently
and can actively regulate itself and its flight on the basis of the data it collects.
First the CanSat is attached to a parachute in a sturdy case. This case is detached together with the parachute
at a height of 200 m and immediately four motors are folded out. From now on the CanSat controls these
motors on its own so the satellite lands autonomously. The deeper it climbs, the slower it sinks and finally lands
gently on the ground. This shows that a drone can also be used as a landing system in space travel and offers
many advantages.
The first part describes the structure of the CanSat and how it works. In the second part the measured results
are presented and in the last part the results are discussed and a conclusion is drawn.
Our research shows that a quadrocopter is suitable as a mini probe for exploration satellites on planets or
moons. With the special salvage system it is possible to fly fast to many points in the environment and to chart
step by step the environment.
The team PerpetuumMobile won the German CanSat competition in 2018 and the prize for the best CanSat
project in the European CanSat competition in 2019.

Keywords: CanSat, drone, landing system, re-landing, remote control, autonomous

CFR-19-PerpetummMobile Page 1 of 22
 1 INTRODUCTION 1.2. Mission objectives
SpaceX exemplifies the dream to us: Affordable space
travel. Since December 22, 2015, the company has come
1.1 Team organisation and roles
much closer to this dream when it managed to land a
The PerpetuumMobile team consists of 6 students and
booster stage of the Falcon 9 rocket again. With the
a mentor from Halle (Saale). We found each other because
successful PR campaign around the Falcon Heavy, SpaceX is
we share the interest in physics, technology and space
well known to everyone. The big goal is to make space
travel.
traveling cheaper by reusing rocket stages and thus
Marius Glaeser (16) Team leader: Marius is interested in
minimize unnecessary scrap.
physics and astronomy. He is responsible for planning and
From the moment we learned about that idea we were
constructing of the CanSat. He is active in all places and
fascinated by the concept and had been enthusiastic about
supports wherever support is needed, so he always knows
its implementation. With our secondary mission, we want
exactly the current status and problems in the individual
to demonstrate that the principle of saving costs and
work area and coordinates the cooperation. 20h/week
protecting the environment can be realized in many ways in
(extracurricular)
space travel and that millions are not always needed.
Cedric Frauendorf (19) Hardware technician: Cedric has
We are now working on a landing system with rotor
a great knowledge in physics and technology as well as
blades as basis to land the CanSat. To do this, first the
experience in model making and cad design. He is
CanSat is attached to a parachute in a solid case. The cover
responsible for the 3d modelling and the hardware design
and the parachute are separated from the CanSat at a
of the CanSat. He, furthermore, is responsible for the
height of 200m and immediately 4 motors are going to fold
cabling of the CanSat and carries out various tests.
out. From this point on, the CanSat controls these
16h/week (extracurricular)
independently, so that the satellite lands autonomously.
Malte Lippert (16) PR+Financial officer: Malte has a very
The deeper it climbs, the slower it sinks and in the end it
good knowledge in physics, experiences in model making
gently lands on the ground.
and in advertising and presentations. That is why he is
This ensures that the CanSat is not going to be destroyed
responsible for finance and public relations in our team. He
during landing and the special landing system allows a
takes care of the recruitment and maintenance of sponsor
further launch. In addition, the landing phase is actively
contacts of our team. He also does the bookkeeping.
controllable that make use capable, for example, to take
8h/week (extracurricular)
countermeasures against wind and drift off. So we are
Fabian Schmidtchen (18) Concept developer: Fabian has
building a mini satellite that lands gently and can actively
a good technical sense for what works and what does not.
regulate itself and its flight based on the data it collects. The
He presents concepts and gives building instructions as well
satellite documents its flight by using a camera and various
as taking care of the radio link of the CanSat. 3h/week
sensors.
(extracurricular)
Oliver Zimmermann (16) Software developer: Oliver
masters C and Python. He develops simulation software for 2 CANSAT DESCRIPTION
the ground station and Arduino sequences for tests.
4h/week (extracurricular) 2.1 Mission outline
Eric Oliver Schmidt (16) Software developer: Eric has a Primary Mission The CanSat measures air pressure and
very good knowledge of Arduino and C. He takes care of temperature. From the measured air pressure, the altitude
programming the flight mode and other CanSat internal and the fall speed are determined. The temperature data
processes. 4h/week (extracurricular) provides a temperature profile of the mission.
Hardy Krause (55) Counsellor: Mr. Krause is a physics Secondary Mission Development and construction of a
and computer science teacher at the Elisabeth-Gymnasium. special landing system for our CanSat. In accordance with
He supports our project. the measured data, the CanSat will land safely, gently and
In general, we do not work at school as we all attend autonomously.
different schools. We base our teamwork on a cloud We use two different systems one after the other for this
system. Everyone is informed about his current tasks by purpose:
Marius. Salvage system: A parachute is attached to a solid casing
that surrounds the CanSat. The CanSat is expected to sink

CFR-19-PerpetummMobile Page 2 of 22
at 9 m/s. When the CanSat is at a height of 200 m, the The CanSat stores all files additionally on different SD cards
CanSat will be separated from the casing and the parachute. to prevent data loss.
Landing system: Immediately, after the separation, 4
landing legs at the bottom and 4 rotors at the top will fold The following goals are aimed at:
out. Now the Quadrocopter steers the last flight phase - Independent, gentle re-landing of the CanSat. Without
autonomously so that it keeps sinking at 8 m/s. In between damage to the CanSat, so it can be restarted.
the CanSat stops to take pictures from the air. It only works - Correctness and reception of all the collected data, as
with the data that it collected during the mission. From a well as the creation of aerial photographs and the
height of 30m it slows down exponentially and lands slowly evaluation of these.
on its landing legs.
Now it records further data. Because of the high capacity 2.2. Structural design
of the batteries, it can take off again independently due to
our special landing system and fly to a new place and take
further aerial photos. This offers the possibility to discover
the surroundings by the CanSat step by step.
Thanks to the special salvage system it is versatile
because it can react to external influences (e.g. wind)
autonomously and actively. It can be used in extreme
weather situations as well. In addition, it is well suited for
exploratory missions because it can be actively be
controlled by the ground station, but also will be able to find
its way on its own if necessary.
Thus, our CanSat essentially is a pre-exploration probe.
First it can travel long distances to the area to be explored
with a rocket and then fly to regional different places using
the landing system. There it sends more and more different
aerial photos and environmental values, which provides an
Drawing 1: electrical components of the CanSat
overview. We are carrying out a technical demonstration of
a special landing system and simulating a pre-exploration The BMP 180 (air temperature and humidity sensor) and
mission in an as yet unknown habitat on earth. In general, the light sensor transmit the measured data to the Arduino
it should send a data packet to our ground station every Pro Micro. The Arduino Pro Micro bundles the data, logs
0.1s. We collect the following values: them on the SD module and forwards them to the flight
controller. The flight controller measures the acceleration,
Measured value sensor alignment, temperature and air pressure. After the GPS and
air pressure barometer (Arduino/Flight compass data have been processed, all data is transferred
Controller) to the transmitter, which sends the data to the ground
air temperature BMP180 (Arduino)/ temperature station receiver.
sensor (flight controller) Parallel to this, the flight controller controls the four
light intensity photoresistor (Arduino)
motors separately on the basis of the measured data, so
Acceleration acceleration sensor (flight
that the CanSat is stabilized. The Arduino Pro Micro controls
(in pitch, roll axis) controller)
the speed of the descent by simulating the remote control.
alignment gyroscope (flight controller)
GPS GPS module (flight controller)
compass compass module (flight controller) 2.3 Mechanical design
aerial photographs digital camera (camera) The CanSat is a hybrid of 3D printed parts and steel
storage location: SD card flight controller/ parts. The basic structure of the CanSat is formed by two
SD card Arduino/ round acrylonitrile butadiene styrene (ABS) plates at the
LIVE transmission to ground station top and bottom. These are stabilized by four supports,
Table 1: measured values and used sensors between each of which a rotor folds out. After numerous
tests, the initial ABS rotors were replaced by a carbon
version, which has proven to be much stronger and more

CFR-19-PerpetummMobile Page 3 of 22
efficient. The cover with spiral outer skin is made of flight status power Capacity
polylactic acid (PLA) and initially surrounds the CanSat. This in in A at used in
holds back the arms and legs. The arms are pressed against Watt 8V mAh
the cover by leg springs and the legs are tensioned by 30 min Standby 1.2 0.15 75
rubber bands. All the important components are threaded 90 s descent on parachute 1.2 0.15 5
onto the steel supports, which are arranged in a 30 mm hole 3 s Stabilization after 170 21.5 20
grid. separation
28 s Sinking flight 110 14 110
7 s Landing 170 21.5 45
209 min Standby 1.2 0.15 530
total 4 hours 785
Table 2: expected power consumption

Available 1500mAh > 785mAh demand

⇒ Condition 4 hours running time fulfilled, 7.5h possible
continuous operation

2.5 Software design

Several microcontrollers are working in parallel in the
probe. In general, the Arduino Pro Micro takes care of the
acquisition of the sensor data and the determination of the
Drawing 2: basic structure of the CanSat necessary thrust of the motors for the respective point in
time. These are transmitted to the flight controller via a
2.4 Electrical design pulse-pause modulation (ppm) signal. The flight controller
then calculates the appropriate speed for each motor from
the transferred values and the alignment (determined via
gyroscope). This simultaneously stabilizes the flight. The
respective speed is transferred to the respective ESC. The
ESC then converts the applied battery voltage of 8 volts into
a kind of three-phase current for the brushless motor.
The Arduino Pro Micro uses the BMP180 barometer to
determine temperature and air pressure. The height above
the switch-on point is determined from the difference in air
pressure between the value at the time of switch-on and
the current value. This means that we are independent of
weather-related changes in air pressure and achieve more
accurate results. The ascent/descent speed is determined
from the difference between two height measurements
taken at different times.
The photoresistor supplies an analog signal value for the
Drawing 3: Circuit diagram of the CanSat
Arduino by a voltage division circuit with a further reference
resistor of 4,7kΩ. Sensors connected to the Arduino are
In the CanSat there is once the 8V circuit, which is
read out every 0.5 seconds and sent to the flight controller.
needed to supply the motors and the electronic speed
In addition, the flight controller determines the position
controller (ESC) with power. The whole control electronics
and orientation with GPS compass and acceleration
runs over a 5V circuit, which is reached by a DC/DC
sensors. All collected data is stored on an SD card and sent
converter. The sensors as well as the microcontrollers are
live over 435Mhz to the ground station.
connected to it.
The flight of the probe is divided into five phases:
Phase 0 (waiting on the ground for take-off): As long as
the probe in the rocket has a height of less than 100m, the

CFR-19-PerpetummMobile Page 4 of 22
motors and propellers are prevented from activating and
folding out. Only when the probe exceeds 100m the
Arduino does switch to phase 1.
Phase 1 (climbing via the rocket): The probe now
measures all data already during the rocket flight. Shortly
after the apex, the probe is released and the descent speed
is evaluated. As soon as this exceeds 5 m/s, the light sensor
checks whether the probe has been ejected by the rocket.
The parachute unfolds and it goes into phase 2.
Phase 2 (parachute flight): The probe flies downwards
with approx. 9 m/s on the parachute. From a defined
altitude (approx. 200m) the load switch is automatically
activated (only if necessary it can be operated by the
ground station), which ensures the separation of the probe
from the parachute. A continuity test determines a
successful separation, the motors are brought to minimum
speed after a small delay (1-3s) and phase 3 is changed.
Phase 3 (controlling of the sink rate): In the Arduino,
depending on the weather situation, the set sink value is
initially set to 10 m/s (bad weather) or 8 m/s. Using a
directly proportional controller, the change in the current
thrust value is calculated from the difference between the
current sink rate and the target value. The peak of the speed
at the beginning provides a large safety distance to the
parachute system and other possible payloads. This
prevents the probe from touching other objects. From a
height of 100m, the target sink rate is continuously
reduced, so that the probe slowly decelerates to 3 m/s at a
height of 30m. Then it switches to phase 4.
Phase 4 (landing): From a height of 30m, the flight
controller now controls the descent speed to approx. 1.5
m/s. The speed of the descent is then controlled by the
flight controller. From 10m the sink rate is further reduced
and the ground effect detection is activated. With the
ground effect less thrust is needed to float just above the
ground than it is a little higher. Because our CanSat notices
that the same sink rate requires less revolutions than Drawing 4: flow chart of the software
before, it can detect the ground effect. If this happens, it
can be applied very gently. The motors are then switched 2.6 Recovery system
off. After ejecting the CanSat, a parachute unfolds (diameter
Programming language: 45cm, spin hole: 10cm). The parachute is attached to the
Flight Controller: C++ (Notepad++ / Visual Studio) cover. The cover itself is fixed on the CanSat by a yarn. At
Arduino: C (Arduino IDE) 200 m this yarn is loosened with a glow wire and the cover
Memory volume: separates from the CanSat.
Camera: 32GB SD-Card (memory for 4h continuous film) The four rotors on the sides fold out (2xCW / 2xCCW).
Arduino: 16GB (memory for all commands and sensor These are guided by a flight controller and independently
values) Flight Controller: 4GB SD-Card (memory for all controlled by the CanSat with the Arduino.
processes) The biggest advantage of a "quadrocopter" as a recovery
system is the active adjustability of the flight or fall speed.
Thus the CanSat will initiate and control a very calm

CFR-19-PerpetummMobile Page 5 of 22
descent for the aerial photographs in order to achieve the We will set up two separate ground stations, each
best possible stabilized image. provided with an antenna. There will be a main station that
The recovery system is designed in such a way that the can intervene in case of emergency and a secondary station
motors (with a battery cell configuration of 2s 2p in total) that only logs the data so that no data loss occurs.
need a kv of 5500 U/V. The propeller has a diameter of 3 If necessary, the CanSat can be controlled via the ground
inches and a gradient of 2 inches. We chose the SunnySky station. A remote control is coupled with the CanSat and
R1104-5500 engine because it is specialized for these can execute all commands (e.g. to control the motors).
propeller types. It is also very light and does not take up
much space.
The folding mechanism is based on a hinge joint. The
motors are folded out using leg springs. This means that the
arms are always under tension when folded in. In order to
prevent premature unfolding, each arm is fixed by the
cover. This cover is printed with a spiralised outer skin. This
prevents getting stuck in the rotors when they come off.

Expected flight time: 130s (90s on parachute | 40s landing)

2.7 Ground support equipment Picture 1: screenshot of the ground station

For the data transmission to the ground station we use
the 3DR-Radio Telemetry. The data are collected by a Yagi
directional antenna and transferred to the PC of the ground Radio frequency:
station. We use the Mission Planner software to record and 435 Mhz (radio link)
display the data. On the left side the flight information and 2.4 Ghz (remote control)
the sensor data are displayed and on the right side the exact
position of the CanSat is shown on a Google Maps map.

CFR-19-PerpetummMobile Page 6 of 22
 3 SCIENTIFIC RESULTS

3.1 General facts

Critical points of mission Time Time from the rocket`s
[hh:mm:ss] launch [s]
CanSat ready in the rocket 17:47:05 --
Rocket´s launch 17:47:33 0
Maximal altitude – 806m 17:47:41 7,5
Full release of the parachute – 670m 17:47:51 17,7
CanSat´s landing 17:48:32 49
Recovery of the CanSat 18:18:38 1865
Table 3: critical points of the Mission

We often included in the diagrams the height for visualization purposes for a better classification of the
measured values. With the exception of chart 3 the x-axis of all charts is the time related to the start of
the CanSat.

3.2 Primary mission

Chart 1: Dependency of the altitude to pressure during the mission. The air pressure is determined with
the BME180 from the arduino and the altitude is calculated using the barometric formula.

CFR-19-PerpetummMobile Page 7 of 22
Chart 2: The change in temperature during the flight.

Chart 3: The dependence of temperature to altitude.

CFR-19-PerpetummMobile Page 8 of 22
Chart 4: The change in temperature measured in a whole measuring range.

Chart 5: The signal strength of the 3DR radio telemetry measured at our ground station.

CFR-19-PerpetummMobile Page 9 of 22
3.3 Secondary mission
1000

800

Start in flight with landing phase

rocket parachute
600
height in m

400

200

0
-10 0 10 20 30 40 50 60

-200
time in s

Chart 6: The flight course: classified into different phases.

Chart 7: The signals transmitted by the Arduino to the flight controller for the corresponding flight phases.

CFR-19-PerpetummMobile Page 10 of 22
Chart 8: The signals transmitted by the flight controller to the ESC for the corresponding motor thrust.

Chart 9: The change in the orientation of the CanSat in Roll and Pitch Axis. (Roll: x-Axis, Pitch: y-Axis)

CFR-19-PerpetummMobile Page 11 of 22
Chart 10: The pwm signals of the remote control of all 7 channels received by the flight controller during
the whole mission.

Chart 11: The pwm signals of the remote control of all 4 (important) channels received by the flight
controller after the whole mission. (re-start of the CanSat by remote control.)

CFR-19-PerpetummMobile Page 12 of 22
Picture 2: The change of the position of the CanSat based on the data of the GPS module. Marked point:
Location of the ground station.

Chart 12: The accuracy of the GPS during the mission. Horizontal and Vertical accuracy as reported by the
GPS module in meters.

CFR-19-PerpetummMobile Page 13 of 22
Picture 3: Map of the surroundings of the take-off zone (1), extracting the area suitable for the landing of
the CanSat (2), mapping the landing zone (3)

CFR-19-PerpetummMobile Page 14 of 22
 4 DISCUSSION The CanSat is ejected 7.5s after take-off and rotates
10.2s until the parachute opens.
Immediately the cord tears, which fastens the parachute
4.1 primary mission with the cover to the CanSat. The knot on the CanSat was
Pressure and height (chart 1) probably knotted too sharply on one side and thus formed
The altitude is indirectly proportional to the air pressure. a predetermined breaking point. That the cord is torn we
It is calculated with the barometric altitude formula. could find out by analyzing the parachute.
The air pressure is very accurate throughout the The cord tears so that the parachute does not pull off
mission, so we can determine the altitude to an accuracy of the cover completely, but only jerks it a little. The cover is
20cm. pushed further upwards by the motors, so that it finally
During the acceleration phase, a back pressure forms in comes off as well.
the rocket, which explains the pressure increase measured Now the CanSat is in free fall. It falls almost constantly
and the negative altitude calculated with it. with a maximum speed of 22 m/s, because the CW value is
very high due to the extended legs and arms. In addition,
Temperature (chart 2, 3, 4) the rotors will have rotated because of the fast draught.
Diagram 4 shows the total time period when the CanSat At a height of 230m the glow wire is annealed, but it has
was switched on. It shows that the sensor is well sensitive no effect.
to temperature changes. You can see significant differences The engines start at 200m and have a considerable
whether the CanSat is placed in the shade or in the sun. The effect at 30m. The speed of the CanSat is too fast, that the
measured temperature corresponds to the actual outside motors cannot brake the CanSat completely.
temperature. But they manage to slow down the CanSat to 8 m/s. So
Shortly before the launch, the CanSat heats up to 43°C that the impact is not so hard.
in the rocket. This temperature remains constant during the The landing legs also intercepted a lot of energy, so the
acceleration phase. When it is ejected, the CanSat measures CanSat didn't suffer any damage.
a significantly lower outside temperature. The temperature
drops to 38 degrees before landing. Now the CanSat lies in Alignment (chart 9)
the field and is heated by the sun. The evaluation of the rotation of the CanSat shows at
In the period of the flight the measured temperature first that the rocket rotates slightly. After ejection the
only applies to the CanSat. The outside temperature CanSat rotates very strongly in all directions. It takes a long
decreases by 4.8 °C to 800m. However, our CanSat and the time until the parachute opens. At 10.2 seconds after the
sensor itself has heated up too much to measure the correct highest point the parachute opens and aligns the CanSat
outside temperature during the flight. After landing and upwards. The parachute and the cover separate at that
before take-off, the measured temperature is the correct moment and the CanSat gets an angular momentum. The
temperature for the outside. free fall accelerates this at first. From a height of 200m the
angular momentum is reduced. The large signal amplitude
Radio connection (chart 5) at the end indicates the landing position of the CanSat.
During the whole mission we had over 75% signal Throughout the entire fall, the CanSat is upright.
reception from the CanSat. This means that we always We cannot say exactly if it is a permanent rotation of the
received all important data packets without interference. CanSat around the yaw axis. However, we strongly suspect
The loss of the connection at 17:49 was intentional. At this behavior due to the periodic and regular deflections of
that time we put down the antenna. Later the connection the pitch and roll values. This will also cause the
was re-established to find the CanSat again. stabilization in the air.
We assume that the parachute has only opened after
We consider the primary mission as successful, all 10.2s, because exactly after this time interval the irrational
objectives set have been achieved. rotation is stopped by a fast impulse upwards.

4.2 secondary mission Motor thrust (chart 8)

Flight course (table 3, chart 6) The flight controller has logged the output signals to the
The CanSat is 17:47:05 in the rocket. ESC. From these we can deduce how the motors have run.
The acceleration phase of the rocket starts at 17:47:33.

CFR-19-PerpetummMobile Page 15 of 22
 Motors 3 and 4 which are exactly opposite and rotate During the acceleration phase we had a very bad GPS
clockwise were controlled amplified. Motors 1 and 2, on the connection and therefore an accuracy of 120 meters.
other hand, rotate counterclockwise. The amplified control However, the accuracy has already improved in the rocket
of motor 3 and 4 amplified was a countermeasure against and thus the accuracy continued to improve during the
the angular momentum. flight.
It is noteworthy that the actions of the motor only have An imprecision is not uncommon with the current GPS
an effect on the fall speed above a height of 30m. standard. The GPS, therefore, is to be judged as successful.
Even if the motors have run only slowly, this would have
reduced the fall speed minimally. Recovery system (picture 2)
We have two explanations for this: Our CanSat was found 40 minutes after landing. The
1) Due to the high fall speed, the stall occurs actual search took only 5 min. (in between there was
immediately and the rotors fall into their own turbulence another rocket launch). We had a good GPS signal so we
and cannot stabilize. Only at higher speeds this status can searched this point with the ground station. The closer we
be overcome. Nevertheless, a small reduction of the fall came to the CanSat, the better our connection became.
speed would have to be measured. From a radius of 20 meters our additional direction finder
2) That is why we assume that the rotors have actually system started to work and we could bridge the inaccuracy
rotated only after 30m, because it was switched to landing of 3 meters of the GPS signal. Additionally the motors and a
mode. Before that the pwm signal is too low, so that the beeper made noise, which were switched on with the
ESC does not let the rotors rotate, which is only the case in remote control.
the landing mode. If the automatic calibration of the ESC The parachute was also equipped with a direction finder
was set too high during set-up at the launch, this may have system, so we could easily locate it.
happened. In retrospect, we find the direction finder transmitter
superfluous on the CanSat. If our GPS signal would have had
Remote control connection (chart 10, 11) failed, the direction finder transmitter would also have
If no signals are received by our receiver, a pwm signal been useless because of the small working radius. The
of 925 will be transmitted automatically. This was not the imprecise GPS rejection was also easily bridged by the
case, so we can assume that we always had a connection beeper.
during the mission. The signal deflections at the beginning However, the systems in cooperation worked perfectly
are due to the pre-launch arming and those at the end are and so we could find the CanSat successfully and quickly.
the result of the searching the CanSat and arming the
beepers. Temperature of the CanSat (chart 4)
The temperatures measured by the CanSat show that
So we never had a loss of contact from the CanSat individual parts of the shell or frame may have warmed up
neither from the remote control nor from the ground to 50 degrees. This value should be viewed critically. The
station. This is to be considered a significant success. base plates of the CanSat are made of acrylonitrile
butadiene styrene (ABS) and the cover of polylactic acid
After landing and finding the CanSat we started it again. (PLA). 3D printing is susceptible to high temperatures. Our
The pwm signals of the remote control can be found in tests have shown that temperatures of 70 degrees make
diagram 11. The CanSat reacted perfectly to all control the case hand-deformable. That's why we paid attention to
commands. Thus it is to be evaluated as completely the temperature of the CanSat and transported it to the
operational and it did not suffer any damage. launch site in a freezer, for example.
It makes more sense to print the CanSat with
GPS (picture 2, chart 12) Polycarbonate (PC) instead of ABS or PLA.
During the whole mission we had a very good GPS signal.
The red lines show possible positions of the CanSat due to Software (chart 7)
the acceleration. The blue line is the actual correct position The CanSat software consists of 2 components:
of the CanSat from the GPS signal, the acceleration and The software on the Arduino has given the right
altitude of the CanSat. commands and stored all important data. Here also the
The accuracy of the position can be read from diagram ejection was recognized and the different flight phases
12. In general we have achieved an accuracy of 3 meters. were switched on accordingly.

CFR-19-PerpetummMobile Page 16 of 22
 The signals from the Arduino to the flight controller can The camera should have run as a shielded system, which
be found in diagram 7. is manually switched on on the ground and records and
The first step is to arm the motors and software. The stores as long as the power supply allows. At this point we
next step is to capture the ejection. At 230 meters the glow tried to save energy and storage space, but this made it too
wire is triggered first and then the signal is given to control complex.
the engines. The last climb tells the flight controller to
switch into the landing mode. The resetting of the signal at Error analysis
the end indicates the ground contact and disarm the Our CanSat has been tested a lot, especially the software
motors. has been tested a couple of times. This was important
The software on the flight controller did not recognize because it has some parameters that we appreciated and
the ejection. As a result, the camera was not triggered, but approximated through testing. For example, how much
there were no further consequences. thrust it has to provide to sink at 8 m/s or how fast the
The signals of the Arduinos were also received, and the motors try to balance without rocking.
corresponding signals were transmitted to the engines. This also means that the software is inflexible. We
However, it was not possible to completely dissipate the considered the hardware to be a fixed factor that always
angular momentum. performs the same way. As with our tests, so too in
The software decides for itself whether it will fly to a Bologna. But now we had a simple hardware error at the
point where the CanSat can land well. But because of the launch and the CanSat was too fast. So the thrust was not
high fall speed an immediately landing was preferred. enough to slow down the CanSat completely. The engine
At a height of 30m it is switched into the landing mode thrust was fixed in a certain value range, found out from the
and the CanSat brakes now with all energy to 8 m/s, tests. The desired speed of 8 m/s can only be reached if the
recognizes that it has landed on the ground and stops the parachute is released correctly at 200m.
rotors immediately. The high speed should have been dissipated earlier by
The software itself had various mechanisms to detect accelerating full throttle. For this we would have had to
whether a sensor was not working properly or whether the install a control mechanism of the fall speed. But the
software was developing errors. However, it was not software had no precautions against hardware failures
recognized that the CanSat sinks too fast. We did not expect built-in. This recklessness was our biggest mistake.
that such a simple hardware problem would occur. It was a
bit gullible that we assumed that the hardware would work Furthermore there was a slight overstrain at the launch.
perfectly. Everyone had a fixed task, but the error of the CanSat was
We should have written control mechanisms for the recognized much too late.
hardware into the software and, above all, carried out more
testing. In order to avoid these errors more tests have to be
done in tricky situations and the software has to be adapted
Camera (picture 3 (1), (2), (3)) by further control mechanisms.
The picture 3 (1) shows the surroundings of the launch
field. We have written a software that analyzes the area and The error that our engines react too late because the
identifies the area where the CanSat can land well. These ESC is not calibrated correctly was known. We always had
images were created by our computer at the start-up. problems calibrating the 4 in 1 ESC. The engines only started
We also tried to record the flight of the CanSat with a to turn at high throttle values. But once they have all turned
camera. However, we have a software bug in the CanSat it fits perfectly again to the remote control and the engines
that caused that the camera took pictures too late. We have also turn at lower speeds. With our software, the ESC must
a 20min video of the grain in the field. always be recalibrated when the CanSat is turned on again.
We would have analyzed the aerial photos with the It does this when it is calibrating its gyro sensors anyway.
software that also extracted the landing area on green The problem did not occur for a few weeks before the start
areas. and we thought we had solved it with the software.
Our software on the flight controller did not recognize Therefore the calibration was not checked during the setup
the apex correctly and did not trigger the camera. for the launch.
This is why the engines only started in our landing mode.

CFR-19-PerpetummMobile Page 17 of 22
 The problem should have been recognized by a simple flight time is limited to only 5 minutes by the capacity of the
test, where the propellers are turned briefly via the remote batteries.
control. After that the calibration could have been tried Altogether the mission went quite well, although the
again. problem with the parachute occurred. Hardware problems
Next time we have to check all problems that have ever of this kind are unprofessional and must be avoided.
occurred before takeoff. Fortunately the mechanics and software worked as planned
during the whole flight.
We never considered that the rope to the parachute is The conclusion is that the mission was successful except
too thin and can tear. We did numerous drop tests, where for the problem with the parachute and fully met our
the parachute's rope was not the center of attention, requirements.
instead we checked if the parachute can pull off the cover. As an outlook this system is very well suited for
We always folded the parachute well and let the system reconnaissance missions if you develop it accordingly and
fall down vertically, so that after 2-3s the parachute was exclude simple hardware problems.
always open. We have never considered that a longer fall On the day we presented our data (28.06.2019) the new
time is a problem. At the launch, however, the speed and project Dragonfly was announced by NASA. It is about an
the force were higher than in our tests and the one knot exploration of the Jupiter moon Titan by a drone-like probe.
was badly fastened. http://dragonfly.jhuapl.edu/
Thus it came to the fatal defect, which could have been More confirmation is probably not necessary :)
eliminated with simple test with extreme loads.
6 SOURCES AND ACKNOWLEDGEMENTS
5 CONCLUSIONS
We have learned all information and knowledge about
Overall, the project was successful and many drone technology with the arducopter system.
conclusions could be drawn from the measured data. Our http://ardupilot.org/copter/ (ArduPilot Dev Team, Access:
goal was to design and develop an autonomous landing 2018- July 2019)
mini-satellite. In order to maximize the reusability and
exploration capabilities of the satellite, the advantages of a We would like to thank all sponsors who supported the
quadcopter were combined with those of a compact project financially or with donations in kind. In particular,
measuring unit to create a unique landing system. we would like to thank DELL Technologies and Telba for
The evaluation of all data showed that an error in the their support.
attachment of the parachute led to a premature
detachment of the parachute from the CanSat. Therefore, Furthermore we would like to say a big thank you to our
the CanSat fell longer and faster than expected without surroundings:
parachute and without started motors. The extent of this To our families, who accepted that the CanSat
problem could be limited by the landing system, although construction had already expanded to five locations in the
we recognized it too late at the ground station. The motors house.
were activated from a height of 30 meters above the To our friends who had to follow the emotional roller
ground. After a short delay caused by the already very high coaster every day.
speed of fall, they delivered enough thrust to contain the To our teachers, especially the physics student council,
fall. Thus the CanSat fell differently than intended, but who have always answered our questions.
could still regulate the flight in such a way that it did not To our supervisor, Mr. Krause, who was always there to
suffer any damage. help and was flexible at every team meeting and made his
After the recovery from the landing field it was possible comments.
to start the CanSat again with the remote control and let it
fly up. This shows that the CanSat is actually directly
reusable, which was our most important concern. It could
also increase the size of its exploration area by flying from
place to place. But for this he has to land on a flat place, Marius Glaeser Halle the 14th of July 2019
which was not given in this case. In addition, his effective Teamleader PerpetuumMobile

CFR-19-PerpetummMobile Page 18 of 22
 APENDIX 29-30.3.19 Wolfsburg Booth: Discussions with 3,000
PR-actions Phaeno visitors and other
technicians
period activity Description persons reached
27-28.3.19 Jugend Presentation stand: Talks 500
21.4.18- Instagram over 70 contributions from 3,840 forscht with visitors and expert
27.9.2019 Channel soldering, technical (national jurors
sketches, tests, paperwork, round)
sponsors...
https://www.instagram.com/p/BvxXclYlDtO/?utm_source=ig_web_b
https://www.instagram.com/cansat.perpetuummobile/ utton_share_sheet

22.5.18- Facebook same contents as Insta 56 16-18.5.19 Fugend Presentation stand: Talks 2,200
27.9.2019 Account forscht with visitors and expert
(federal jurors
https://www.facebook.com/CanSat.PerpetuumMobile/ round)

25.1.19- Twitter same contents as Insta. 32 https://www.instagram.com/p/Bumm24onlr-/

27.9.2019 Account Occasional collaborations
with CanSat_de
2-13.2.19 Kindergarten Rocket construction with 5 23
https://twitter.com/CansatP?lang=de Halle Dölau year old children

03.01.2019- website All information about our project. https://kindelternzentrum.de/

27.9.2019 Links in all social media
11.03.2019 TGZ Presentation for employees 12
https://cansat-perpetuummobile.de/
https://www.instagram.com/p/BvzxwSsHIdc/?utm_source=ig_web_b
17.11.2018 Open Day of Information stand for 50 utton_share_sheet
the ELG parents and pupils
12.04.2019 ELG parents' Information stand for 55
16.06.2018 MINT-fair Exhibition stand: Talks with 1,120 consultation parents and pupils
visitors | Brewing: Talks day
with youtubians and
politicians 12.04.2019 radio Corax Radio report about the 4,200
project, live and online
https://www.instagram.com/p/BkGY6XdhliY/?utm_source=ig_web_b
utton_share_sheet https://kira.radiocorax.de/letzter-schultag-social-lounge-und-
satelliten/
28.6.-1.7.18 Silbersalz- Trade fair stand: Talks with 1,500
festival Germany's largest science 21.05.2019 Mdr Television report about the 60,00
young talent project 0

https://www.instagram.com/p/BkpQHo2HScP/?utm_source=ig_web_ https://www.mdr.de/mediathek/mdr-videos/a/video-305300.html
button_share_sheet

21.08.2018 Eigenbauko informative lecture 14 21.09.2019 AufDistanz Podcast about CanSat with 1,000
mbinat Halle interview from us

25.08.2018 Posters in Information about the 925 http://aufdistanz.de/auf-distanz-0044-der-deutsche-cansat-

ELG project for pupils wettbewerb-2018/

https://www.instagram.com/p/Bm5sMm_nkLy/?utm_source=ig_web 09.07.2019 Mdr Television report about the 90,00

_button_share_sheet project and launch +radio 0
report
06.09.2018 student Guest article about the 720
newspaper project https://www.instagram.com/p/Bzs8xHOnvuB/

https://elize-halle.de/satelliten-bauen-bis-es-raucht/ https://www.instagram.com/tv/Bz0H0sznPKH/?utm_source=ig_web_
button_share_sheet

CFR-19-PerpetummMobile Page 19 of 22
Time schedule

CFR-19-PerpetummMobile Page 20 of 22
CFR-19-PerpetummMobile Page 21 of 22
Flyer of the project

CFR-19-PerpetummMobile Page 22 of 22