MSS - Software for planning research aircraft missions
This is a modal window.
Das Video konnte nicht geladen werden, da entweder ein Server- oder Netzwerkfehler auftrat oder das Format nicht unterstützt wird.
Formale Metadaten
| Titel |
| |
| Serientitel | ||
| Anzahl der Teile | 160 | |
| Autor | ||
| Lizenz | CC-Namensnennung - keine kommerzielle Nutzung - Weitergabe unter gleichen Bedingungen 3.0 Unported: Sie dürfen das Werk bzw. den Inhalt zu jedem legalen und nicht-kommerziellen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen und das Werk bzw. diesen Inhalt auch in veränderter Form nur unter den Bedingungen dieser Lizenz weitergeben | |
| Identifikatoren | 10.5446/33790 (DOI) | |
| Herausgeber | ||
| Erscheinungsjahr | ||
| Sprache |
Inhaltliche Metadaten
| Fachgebiet | ||
| Genre | ||
| Abstract |
|
EuroPython 201735 / 160
10
14
17
19
21
32
37
39
40
41
43
46
54
57
70
73
85
89
92
95
98
99
102
103
108
113
114
115
119
121
122
130
135
136
141
142
143
146
149
153
157
158
00:00
SoftwareFormale GrammatikRechenschieberMultiplikationsoperatorAutomatische HandlungsplanungResultanteWhiteboardTwitter <Softwareplattform>ComputeranimationVorlesung/Konferenz
00:45
Prozess <Informatik>Prozess <Informatik>PunktwolkeBildschirmfensterAbgeschlossene Menge
01:19
Prozess <Informatik>PrognoseverfahrenEinflussgrößeSimulationParametersystemVarietät <Mathematik>ModelltheorieSatellitensystemSweep-AlgorithmusPhysikalisches SystemGruppenoperationKollaboration <Informatik>Spannweite <Stochastik>Wurm <Informatik>Lokales MinimumAutomatische HandlungsplanungAdressraumGlobale OptimierungDimensionsanalyseDatenmodellNebenbedingungIterationDifferenteGenerator <Informatik>Heegaard-ZerlegungBeobachtungsstudieComputersimulationInformationEinflussgrößeRechenbuchStreaming <Kommunikationstechnik>Mechanismus-Design-TheorieGruppenoperationFaktor <Algebra>Prozess <Informatik>PolstelleMailing-ListeBAYESMereologieFokalpunktSatellitensystemPunktwolkeDatenflussMomentenproblemSpannweite <Stochastik>Kollaboration <Informatik>Algorithmische ProgrammierspracheMultiplikationsoperatorDimensionsanalyseBitEndliche ModelltheorieVorhersagbarkeitModelltheorieStützpunkt <Mathematik>WhiteboardInstallation <Informatik>SchnittmengeÄußere Algebra eines ModulsQuick-SortArithmetisches MittelOffice-PaketMAPDialektGesetz <Physik>Ordnung <Mathematik>Open SourceDickeSchnitt <Mathematik>PunktSoftwaretestHalbleiterspeicherStereometrieKlasse <Mathematik>Rauschen
08:54
Sweep-AlgorithmusUnterstützungssystem <Informatik>Automatische HandlungsplanungSoftwareClientServerVersionsverwaltungNominalskaliertes MerkmalModelltheorieStandardabweichungSpezialrechnerURLKomponente <Software>Graphische BenutzeroberflächeVertikaleGerichtete MengeRelation <Informatik>PrognoseverfahrenPunktInformationArchitektur <Informatik>BenutzeroberflächeClientOpen SourceAggregatzustandSchnittmengeGruppenoperationVorhersagbarkeitDatentypKartesische KoordinatenPunktServerSoftwareMultiplikationsoperatorRechenzentrumSupercomputerQuick-SortBitrateEinfach zusammenhängender RaumKomponente <Software>DifferenteTrajektorie <Kinematik>EinflussgrößeBildgebendes VerfahrenSystemplattformInformationRechenschieberWort <Informatik>MAPWeb SitePhysikalisches SystemDimensionsanalyseKugelkappeOffice-PaketBaumechanikMapping <Computergraphik>Klasse <Mathematik>Natürliche ZahlLie-GruppeStörungstheorieMomentenproblemSchlüsselverwaltungProjektive EbeneMathematikProzess <Informatik>Ordnung <Mathematik>TabelleMusterspracheTreiber <Programm>Endliche ModelltheorieOffene MengeGrenzschichtablösungBenutzerbeteiligungIterationSechseckSoftwareentwicklerStützpunkt <Mathematik>ComputerarchitekturOverlay-NetzDebugging
16:30
ClientServerOpen SourceInstallation <Informatik>Graphische BenutzeroberflächeDefaultAxonometriePunktProxy ServerMailing-ListeLoginURLSweep-AlgorithmusWidgetTabelleVertikaleLaufwerk <Datentechnik>SatellitensystemOverlay-NetzSechseckMusterspracheCodeKreisflächeOffene MengePhysikalisches SystemFormale SpracheGeradeTotal <Mathematik>MAPBitMultiplikationsoperatorMathematikTrajektorie <Kinematik>DatenstrukturWeg <Topologie>CodeSchlüsselverwaltungGeradeStellenringDialektCachingOverlay-NetzSatellitensystemp-BlockSoftwareInterface <Schaltung>SoftwareentwicklerProxy ServerDifferenteSpannweite <Stochastik>PunktZeitrichtungKartesische KoordinatenEinfach zusammenhängender RaumInformationProjektive EbeneOpen SourceEndliche ModelltheorieGruppenoperationExpertensystemMinimumAnalysisRechenschieberSharewareQuantorServerElektronische PublikationKonfigurationsraumBitrateDokumentenserverPlug inSchreib-Lese-KopfEinfügungsdämpfungMereologieKette <Mathematik>t-TestSoftwarepiraterieProgrammierumgebungZellularer AutomatBereichsschätzungModelltheorieRadikal <Mathematik>SprachsyntheseBildschirmmaskesinc-FunktionAutomatische HandlungsplanungGesetz <Physik>Papierloses BüroPhysikalisches SystemPhasenumwandlungDeskriptive StatistikWeb SiteGamecontrollerNormalvektorAggregatzustand
24:06
Vorlesung/Konferenz
24:29
SoftwareAutomatische HandlungsplanungDatenfeldVorlesung/Konferenz
24:52
Offene MengePhysikalisches SystemFormale SpracheGeradeTotal <Mathematik>SoftwareTaskWort <Informatik>Office-PaketSchnittmengeVorlesung/Konferenz
25:59
Rechter WinkelVorlesung/Konferenz
26:24
Basis <Mathematik>TypentheorieEndliche ModelltheoriePhysikalisches SystemComputeranimationVorlesung/Konferenz
26:45
RückkopplungKonstruktor <Informatik>BitrateApp <Programm>Vollständiger VerbandVorlesung/Konferenz
Transkript: Englisch(automatisch erzeugt)
00:07
Yeah, thank you. And I have an overview slide about me. My Twitter handle is like my name without the blank. I'm working at an institute for atmospheric science
00:23
on the Forschungsten Jülich. I became, some years ago, a fellow member of the Python Software Foundation. I'm also a board member of the German Python Software Band. So, and I have to do something for a living, and this time I have to talk about my work in the institute.
00:46
The institute is doing research on atmospheric processes. If you, I think the most of you there know how clouds look like from aircraft flights to holiday vacation, something.
01:01
And there are a lot of processes to look at through a window to see the atmosphere, but not all of this is understood. Otherwise, we would have a minute forecast of weather forecast, which is perfect. But we don't have, so there must be a lot of things done.
01:20
You may remember a little bit from school, or I don't know, some may be also more interested in this stuff. So we're living in the troposphere. This is the region up to mostly 10 kilometers. And there, if we have some of this biomass burning
01:41
or some of the factories that produce some kind of substances, also we can smell. So, and these two are also reactions in the atmosphere. And then you may know thunderstorms, the lightning that also produce reactions.
02:02
And in the atmosphere, there's transport mechanisms. And a lot of this is already understood. But we always find with some calculations that we have some gaps in our knowledge. Then the atmosphere also has transport from the tropes to the poles. And at the poles, you may know
02:21
from previous documentation that there is a ozone hole at the poles and so on. And above this troposphere, there's another region that's named stratosphere. And above the stratosphere, there's also another region that's close up to 80 kilometers or so. And yeah, there is some transport mechanism
02:43
which go from the troposphere to the stratosphere. And the aircraft, they usually don't be able to get in this high altitude, which is the stratosphere. But there are also some aircrafts we can hire which can do this. And we do then process studies in this atmosphere.
03:05
And this is mostly a focus because we want to know more about the climate. We are all, you have probably seen on Twitter, there was information that there is a big ice
03:21
splitting up from the Arctic ice. So that's because of global warming and that's a problem for our next generation. And in the past, all again, question already has a problem with the ozone hole. And so you can do, your knowledge can be based on laboratory studies so that you can have
03:42
the actions in the laboratory. And so you know some of the actions take some time and you can then find out which are the substances and then you need instruments which can measure the substances in the atmosphere. So you can use balloon experiments, aircraft experiments,
04:03
or also satellite experiments and all groups doing these experiments, they work together. And with this data, at the moment you know some reactions, how the flow in the atmosphere is, you can also try to make some simulations
04:20
and then you can predict some of the measurements you will do and if you measure these values, then you know that the prediction was right. Okay, our most goal is we want to understand more
04:41
how all these processes in the atmosphere work. This can be very too atomic processes, that's their reactions, that's their transports, how clouds move to some places. And this can be parametrized and then this gets into different models. And with all the measurements,
05:01
you also can help to improve all the models. So then you get better forecasts or better forecasts for some problems, maybe if the industry produce more hydrogen, you will also want also to know if that has some effect.
05:21
And because the problem is how to get to this data, we do this by aircrafts because it's cheaper to hire an aircraft as to build a satellite. Of course, you can get more data than by the satellite,
05:41
but you can go to some places now where you need 10 years, well, we have 10 years later of a satellite because the procedure to build up this takes very long. But at all, if you hire an aircraft, does that not mean that it's very cheap? It's not like 180 euros to Bologna,
06:03
it's much more expensive. And so the scientific groups in Germany and the Europe, they work together and they hire in collaboration in aircraft. So, and there are various groups, they are specialized in atmospheric research
06:21
and they build instruments which are automated and they can be added then to such an aircraft. For example, to get into stratosphere, we can hire a Russian spy aircraft, which was a spy aircraft in the past and it got an altitude about 20 kilometers
06:41
and but has a range of 3,000 kilometer. And on this aircraft, you see a lot of slots or bays where instruments can be added and you can imagine that there are many groups of this aircraft and they want to use the aircraft and have some scientific goals
07:02
and you have to get them all together because everyone paid some money for this aircraft for some weeks. So we can also add to wing parts or to the major bays some stuff. That's an example list shortened, it's very long.
07:22
That's a lot of instruments, group names for substances which can be added to this aircraft and depending on the mission we can make, then the instruments can be chosen. Another aircraft is the Harlow, it's a gas stream.
07:45
So the difference between the Harlow and the Geofusica is that the Geofusica has only one pilot, there's no one on board to manage some of the instruments and the Harlow there can be scientists
08:00
within the instruments and can control them. Another difference is also the range, it can be 10,000 kilometers and up to 15 kilometers. This picture shows one of the bases we use in Kiruna in Sweden. Now you have these groups, you have an aircraft
08:22
and now the problems can start because they are from different groups and you have to manage to get all their ideas together and you can use model calculations to make some prediction of the atmosphere
08:42
at the time you may have a flight. So the problem is that in, usually we always think it's three dimensions but that's not right, you always have four dimensions.
09:02
You have always longitude, latitude, you also have an altitude and then it's in, for measurements also the substance of where you are interested in, this has also a value. So for example in some height,
09:20
you have an ozone value of this amount. And so you can calculate this of the model data and amounts of layers which gives you information
09:41
how it will look like if you fly there. And then if you have put all these ideas in your mind together, you also have the problems that aircrafts cannot fly everywhere. So there's over flight problems, so it's not, there may be a military base
10:01
you have no idea from because you're a scientist and then the pilot says no, I'm not allowed to fly there. So there's a lot of iterations and discussions which usually could take long time but this was the reason that these groups have decided to develop different kinds
10:24
of mission support systems. So this ideas to get together started in 2012 and many groups have had the same idea and the mission support system by Mark Houtenhaus was a successor because he was able to use, to develop it in the DLR
10:46
which was hosting aircrafts and he was able to design it from the beginning as open source project and really an open source project becomes in 2016.
11:02
And now a lot of groups are working with the software and developing it and we now have also a future rise to Python 3 but it's built for Python 2 and it's a client server application. The software also uses standard,
11:21
it's an open view special consortium web map service. You may know this from the GIS server. The idea from Mark Houtenhaus, we use the same idea as all the mapping servers have. We calculate our data on a map and then we can use all these tools.
11:43
This software is also based on Anaconda because it has to run on several platforms, Windows, Linux, Mac, and it's easier to have it instead on your own, generally we are going to Connor Forge
12:01
so it's a Connor Forge community application now. License of actually two and lots of documentation on RTFD. Mark did also in this because his idea was so quite good, he did a publication in GMT describing his idea
12:23
to use a web-based service to plan all these flights. And the principle of this web map service is you have a server with a big amount of petabyte, terabyte of data which you have accessed from some cluster
12:43
or some supercomputer which can handle this data and you don't have to transfer the data to the airport where all the discussion with the pilot have to be done. So and you have a web map service request by a client and then to the server about some region
13:04
and also some elevation and some substances and this server makes an image and sends this back to the client. So you can fast scroll by a client with a very poor network connection to a big amount of data.
13:22
And you don't have to transfer the data just only the images to the client. So what we now can have by this is the MSS is at the same time a server and also the client application
13:41
so that the data center can install the server component. If someone has other data than already the implemented one, then he can add some methods to it and handle its data. We have now that's ECMWF data sets
14:01
for its chemical non-conjure model of Ulyss and gravity wave forecast data and some material set data and so on. They can be just driven by the server. The client is secured four or five application which then accesses this server.
14:21
It can also access other existing servers and overlay or look at these data. And this is just done by an HTTP request and so that's now the front end for the scientist. And in this front end the scientist can draw a flight path.
14:42
And all the time he changes the elevation or the layer or the time, forecast time, this flight path stays on top and he can just focus on the data he has from the predictions
15:02
where maybe interesting spots where the aircraft should go. And this flight path is always overlaid on any view. And of course you can exchange flight paths so that you can have with different groups
15:21
look at these data. So the server has the architecture is that it is a risky application and it's based on this standard. And you have to give data driver to read your data and also a vitalization methods
15:40
so for horizontal or vertical views. So that's what the server then delivers to the client. The client has currently major views. It's about a top view and a side view for the vertical stuff. And also it has a table
16:01
within all the flight points which also can be described by at once edit and different other flight patterns. There are some instruments they want to have a hexagon flight pattern on those that can be added there.
16:21
And all views work together and gives them the flight tracker. Installation is quite easy as it is a CONA Forge application. You have to edit the CONA Forge channel if you don't have, create an environment,
16:40
activate it, install it from CONA. And we have all the server data implemented so that you can set up a local host, demo data server so we can check out how this at all works. And on the right side you see the starter launcher of the JUT application and on the bottom
17:03
how the VSCII server stand alone runs. There are a lot of configurations done which could be overwritten by a tracing file. For example, you can enter different map projections or also write your own import and export plugins
17:23
or if you have a very, very slow network connection you can add a local block proxy and switch the caches and all these files. The top view looks like this. This shows the blue line here.
17:42
This is a flight path. Also edit with some description. And you have the usual interfaces of Matplotlib. You can zoom or pan inside. You have the possibility to change layers.
18:02
There's an auto update feature. Also keystrokes are supported if you use the arrow keys or the enter keys and you have not to use all times more. So this is a view just on the map. If you have this flight path on side view
18:22
you can also by mouse move the points in altitude and look at the table or if you have data about the fuel consummation of the aircraft you can also add performance data of the aircraft so you know about the range the aircraft can do.
18:45
And if you look this up with data you can always see for different kind of speeds in the vertical view, this is above, how which region the aircraft will pass
19:03
of predicted data and then this gives the information for which instruments this may become interesting. And on the top view you see this is a map which looks a bit like a weather forecast map
19:21
you may know from TV. So and you can also have also the flight path here as you see which region are then over flight. The thing also has add-ons. For example, if you want to compare with satellites
19:42
you should have overpass of a satellite where you measure so you can access satellite track. Then there's also for some removed sensing add-on and one of the newer features is a KML overlay.
20:01
Sometimes it's interesting to know where are the borders or where some spots you have to go and then you can access via KML. This shows an example if we did a flight
20:23
from Kiruna to Oberpfaffenhofen. This is on the left side, it's a substance which is used for analysis, it's N2O. And there are some regions where this substance changes.
20:44
They are passed by the aircraft and the pilot has added all these waypoints. Pilot has to pass some waypoints which are signals for him. And on this side, on the right side,
21:02
you have also a mixing ratio. So it also shows just we would pass in this region if we fly this way. And in the vertical views you see what this oil model data what kind of structure or what kind of changes
21:23
in the atmosphere the aircraft will pass and where some data will be sampled. There's a lot of campaigns. Campaigns means that our groups hire an aircraft and build together at some places.
21:41
Then in 2014 this was, this is in the past of course, there was some from Oberpfaffenhofen. Then 2016 was Polestag, it was from Kiruna. Then actually there's this TretoKlim campaign that's a new project which is just started from Nepal.
22:01
So this Russian aircraft will be in Nepal and this will fly over India. And so then also Nordex was in Iceland, Emerge is currently running, it's in Europe and Asia and the Vice campaign will be in Ireland. So there are many more groups which will use
22:21
this software to blend their flights. We have documentation on RTFD. Also these papers, we have a repository on Bitbucket. So that's also the main development done. Also that is now, that I'm now on Fortune syndrome
22:41
did not change, that it's an open source software. I was agreeing to maintain the software because it become open source so we don't have the issue that some other company now it's the same as some that's don't have problems maybe with open source. So it is this open source, it can't be changed.
23:03
All we live in an ecosystem, I want to mention this too. We can do such great softwares because us has also make great software and they are also free for us. It's a Bitbucket, GitHub, PyCharm, Quantified Code which unfortunately will be closed in July,
23:22
PyTest, PyP, Anaconda, the good, great CondaForge community, a lot of tools, Jenkins, Travis, Avayo, CircleCI, and of course, Reads the Docs, it's all these tools or experts system
23:40
would not so great and it won't have so great tools. So that's my last slide about the mission support system. We are also added to open hub so we have some more analysis. The last release is below 20,000 lines of code and mostly done in Python and we have active contributor rate of eight.
24:02
Thank you. Questions? Oh yeah, now it's working. Anyone any questions for Raima?
24:27
Okay, so thank you very much for your talk. This is not a question directly related to the talk but are you aware that there is flight planning software also for Sofia which is the big 747 aircraft that is used for space missions
24:40
and have you had any contact with these people? Yeah, the question was, are there other flight planning software? I don't know if in a sense that's right.
25:02
Sofia, no, although we use these conferences
25:22
and also other documentation and also campaigns to spread the word that this software is free. We like to have contributors and we like to have it extended. So, although it's a task to you, if you know those people, tell about the software
25:41
and we are happy if there are more contributors and it gets a bigger scope. Okay, another question? So, oh yeah, Sofia. I was just wondering, in the beginning you said
26:01
that researchers rent this airplane, right? But from whom they actually rented? Rent now? Eh, sorry? No, I don't understand. Ah.
26:24
Yes, it's an MBB, it's hosting the Russian aircraft and it can be rented there for science. So, and the Harlow is based on the DLR and you can also rent it there.
26:41
And there are many others. The DLR also has a Falcon aircraft and there's another company for Envis Copis has also some access to aircrafts but you have to find a company which gives you access to the aircraft. Yeah, if not able to get military aircrafts
27:01
like drones or something like this. Okay, so please go to the app and rate the talk or give constructive feedback. Thank you very much, Reimer Bauer.