Bestand wählen
Merken

Tempus: A framework for multimodal trip planning

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
this so I again and you go from the sun we are as small as a French company did orange interim GIS Open Source I'm going so templates which is a framework for this single tree flooding in the context of these projects and so it's a platform for mutual route planning that means that we are interesting in lots of of of transport transport modes and will the proof treats area under ship dates is true is that's a collaboration with the research labs uh mainly in the style and so mind in France and the main objectives are the first one is extensibility of the framework so that we can have a and address lots of different scenarios and also performances and the and the targeted use cases are we want to be able to develop new moon route planning algorithms so it's more research oriented uh developments and I have a of a common framework to tests Zam into benchmark ramps and also would like to have some predictions k-dimensional route planner right pretty young projects have the 1st open source for me is dates from the last month of May in 2 14 um and I'd be remaining mostly the the only 1 working on that's us I'm gonna give you an overview of the the current much data of the projects of Tempest so is that we have as CP at the core of of the the project suppress breast the that is responsible of holding in memory in the graph and they can expose its API to a certain amount of planes uh so each routes under rhythms is implemented in a in a plane in and now it's also in the cellar that is that's used but is the WPS protocol to communicate with the external obligations we have all the data come from a database and apologize that amazes particle or schema so in the in the database you have rows and older transport offers and in addition to that so temperature ships with the with that so there which will which is made in in Python and fuel so you can loads external data sources is of course from OpenStreetMap and also for the of parts from the typhus feeds off from all the time kind of data sources and on the other hand on the other hand sorry um region of client we we have client application that can request the court and the main is a picture displayed in that is being developed a specifically for Tempest and we also have a Python API that can be used 2 2 rights batch streets and all benchmarking scripts at 2 to test algorithms so I'm going to detail on the internet in the midst of those components the then the dentist core um stores all the graph in memory it's so it's written in C + + chooses boost graph and but and all the graph is loaded from a post here is that the base it's able to answer as a WPS requests and it's multithreaded so you can answer as in the request in parallel is is designed to be as lightweight as possible the but you have a particle draft that's a modern uh that's we call national rough means so and the the the root of the model is the road network and on this 1 into which again i connects all the graphs and namely that the the graphs that can be transported graphs and you can also connects point of points of interest like parking is shared by stations should count stations all and you link all the layers and on the road pages so that's you of the different graphs and you can also enable or disable for instance song at public transport graph for about 2 propagation if she wants and when this particle model we have time so we we we needed to to to design a special there special editor and distance rooms because we have some kind of a meter graph on top of the other graphs on top of the real graph and the predictions for graphs and it's seen as it was draft so it's respect the the simplest most consent of was graph each node is is a is a node of the underlying paragraphs birds for edges it's a little bit different so we have trespassed special rooms so I'm sorry for my lack of artistic scales but so we have here and is risk in its representation of and so we have a tool road nodes to the b transports nodes that that are attached to a room and age and GUI and that is also attached to it to revenge to a different kind of and it's just FIL so this 1 is from from a road style from around node to a remote we have that rule to GUI GOI into roads road to to coming to sponsor station and for me just what this station to road tend to be transported from 2 to 4 so that's we can see all the nodes in one's 1 honoring graph with this particular uh I just since you work so this was the corn and without reliable to design plug-ins enriching buttons and of apart from the simple but is not the ship to is we we over and and not sorry so it should it's supposed to splitting is the is did is declared as a as loadable modules so it works knowledge and the Linux and Windows as adult this award of yellow and so it's really basic foragers plugging that can be a vector that can be requested by the user and then receives the global graph inputs and they must produce some and a certain amount of fluid maps when it's 1 remote but if because we we want to be able to and certainty objective and problems you can
also findings can also produce Malvern 1 another and it allows to reason experiment new features so the idea is to to achieve the course simple and when the plan is become too complicated to try to the ports features from phonemes to the to the core and the main factor in sectors plugin so it's so in terms of logarithms pretty simple it's a it's a store but we we we focused on on this and things that so for instance we are already able to mix different transport modes like she wants to be walking used by the transporter also use should bison shut cars if you if you like and also at the aware of markings for instance if you want to go from point a to point B with your call you want to wish to biological before arriving at the destination but we also are able to handle 10 restrictions and uh more recently the speed profiles uh so if you if you have some tracing that's on some roads age is you can you can be at taking into account the photon restrictions and this is a very generic term and concept that is a is in place the and it's able to take into account so possibly complex demerits restrictions and full fry instance so you have this you can do a U-turn here but so is is not just a 2 2 segments you have to pass through some different from edges and it's so it's of original richer in way of dealing with that's it's use is the internal and automated a state mission for each restriction about speed for fries and so on if you update on you can lower and daily speed profiles on on the on some of the ages and that the engine was taken with the take them into account so if you know that so I know for lunchtime this part of the road is the advance you can you can use this information if she was about the the set of Python scripts that used to load data into the bridges that the bears conveys so for food so we have different early and load is of course for the OpenStreetMap a case and also for other proprietary formats enough they called on them and region it's an we also able to loads elevation data from the city and more in the entity which is typically French the way fall for the pending transport offers of course reality to do with Virgin TFS and standards as for the POI there's no real standards for POI so we we use a basic shape fires we're going with a small language language which quotes uh where we can extract data from from finds and envelopments so there is these things from the IGN inference of so for the French road network and also for public transport this is also from something French I think but anyway with so we are we are developing in this turn so just to give you an overview of all the the the loading process so we have this slowed tempers Python scripts and then you you type the your input so for instance here that I suppose I am ashamed fighting supportive of 1 OpenStreetMap area here and around the city of Smolensk inference what I say OK I want to load with the type of OpenStreetMap my dates are located here and idea the the connection the connection string to the budget is that the Bayes and that sits and if you want to use native a PDF file of form of OpenStreetMap that's as to possible that's needs intimate of processing through osmosis to to converted to XML and then we we we provide the small screen small executable that will translate this this it's similar to should fight and then the chip find means good loaded into into budgets it's really imports time restriction that are declining in OpenStreetMap so there are few amendments it will use them and it's it's real so and the topology processing and I don't know if you know that uh the opposite of that is that the data and that you you found in the press about on directly usable for reading if you have to do with a bit of postprocessing so for instance you can have to if tool to roads that cross a trauma so visually there that cross meant um but so you don't have any of the nodes 1 on the cross points so it's not OK for a for a draft form roughly have 2 2 2 splits sometimes after spins rules but you have to be to take care of have to be aware of some special cases where the 2 the 2 worlds are really not crossing because from for instance you could have that you know for a bridge so you don't want to spend so it in those states for but OK just another view of POI imports of so from a ship find and the the only thing to note is that you can the users so small filter printed here so from this trip finally wants to extract data with that with the current time which it was once it was with me you want and and so on but OK and the last 1 of the components the the WPS summer so for those who don't know a WPA stand for when processing the but it's it's from the OGC and it's basically a based MPC remote-procedure-call person you recall and its use is 1st CGI attendants multithreaded so you can can process the request in parliament and is based on XML so input units and the user requests and you you have to to bad to 2 answers with them to answer we with roadmaps as when it's OK so how about clients so we have a Python client on top of the WP as clients and from that we build the vegetated changes planning it also set for a unit tests and form a batch streets so let me give you some men the inside of the the judges plane this a is the the
Wilkinson's changes of these OK this is that the tend here uh so the different tabs yielded from screens we have me to connect to the cellar and do my requesting and there's a bike and I also have some fancy tools here where I can node direct city low dates are straight from the budget is that's about as an into it also and style sheets so you can see this again they're going we can have the good news of his feedback when you when you divide your your algorithm then here and I connect to the to my w Pissarro and I of the stuff the valuable planning so here I'm going to use the dynamic mentor planning and everything in here below our options that's odd detailed by depending so i'd every funding developer can add its own set of functions and it's would be ultimately automatically generated here so and then this is my Crary uh time of sorry of I can have lots of destinations but also footprint area and there are lots of transport modes I can use so I'm going to start with
the something from point a to point B with the kind of the it was with the private come I think this so I'm allowed to to go by will tend to use my credit card I read this I would have to park the car before arriving when it's when it's being computed and then I ever and idle if it's really clear but on the maps you can see the results here in different colors each color always a different transport modes of many of the resulting roadmap that skin display here and you can again selects his subparts of the of the road map and here there is a mood change so this is my private common that I have to park like on the floor I go to my final destination but and because they have alleviation dates I in my in my that's amazing KIT and also if I want to display them another query I think the yeah so ii enabled in the is the option where I want my private lake and this nation and then into and difference the case I am just using my private car result overall mark ch
example we've got shut bytes story of layers that allows me to display where are located uh by sharing stations the school here OK is the residents so mean yeah In think this is my admired by H and in red this is me walking so I'm I take a bite here if good bye bye and then uh drove the bike but OK and so on and then of course when your your biking the the elimination of profile is important interesting but
just the just to illustrate the the section but so as well I can I can download a layer that represents the some restrictions here in so basically that's a U-turn restrictions and I want to go by car from what I want to do a return to make internal in the then it will find their way around OK but
in the last 1 with public transports so here and we use only introns knowledge and of course because it's a public transport query I have to to specify a time of departure In all is the results OK there were a lot of OK and something that has been added essentially people in the there's been
an and recently is an the image into reverse the graph so that we can solve queries like i want to arrive before a certain time so it's like you want to go after but with the graph reversed so now you can say i want to arrive before this time and it will give you the answer so have the form this idea and something but 1st feature of it is in this printings video sort yeah so it has the ability to produce a trace so the algorithm can in In addition to the to the roadmap they can also uh sense um part of the graph that has been druggist so here it's so we know the the standard logarithm was the ditch and I can I can I can compare with the based
results going some going to enable this option and units and I have another trace which is smaller because I'm using an heuristic for that and the like and I can compare the 2 the 2 traces that's very useful when you develop a reading algorithm OK to so I kids
just a few of us performance is an OK just to its you know we will we what we know we are Swiss there's 2 rooms for optimization in them that similar case but at the same time it's a hard the case but have pretty good results with that and this so 1st tests um 1 good question this scale in some of memory we are originally working with the small French regions were well where everything works OK there we are working on on the on some of some been here scale like you the state of the size of a country of or you wrote 1 or more and some of CPU because we are using as much as 20 and the summary it's k Morris let line only if it's pretty would and all for the future of I we would like to integrate more algorithms like here for of pure public transports is rocked or on the 1 with which is the main condition scan for the role that the case of the Iraqi people rule metaphor and especially the 1 Wankel construction your and which we already know that it's hard to adapt to the dynamic nature of the material there's a small problem we like to into a new data like for instance uh roads that are free from my advice of for for a certain transport mode and the gets uh axis data uh and and use them and then integrate you transport Moore's lecture elegiacal rituals uh where you have to deal with charged discharged station and so on and you optimization criterium the simple the simplest path the most confortable 1 and and the minimum with the with the minimal costs we don't have anything related to to fails on so far and of course find cigarettes I mean the objective of musicians and in the and that's the goal and the connected so if you want the the project is published his home this is determined by the visitor is so if you want to other can contribute to welcome thank you part people
Punkt
Ausbreitungsfunktion
Adressraum
Ungerichteter Graph
Homepage
Internetworking
Eins
Netzwerktopologie
Client
Unimodale Verteilung
Algorithmus
Prognoseverfahren
Bildschirmfenster
Speicherabzug
Meter
Skript <Programm>
Wurzel <Mathematik>
Benchmark
Softwaretest
DoS-Attacke
Addition
Zentrische Streckung
Datennetz
Datenhaltung
Template
Quellcode
Ein-Ausgabe
Kontextbezogenes System
Arithmetisches Mittel
Texteditor
Kollaboration <Informatik>
Rechter Winkel
Beweistheorie
Festspeicher
Heegaard-Zerlegung
Projektive Ebene
Benutzerführung
Instantiierung
Hauptreihe <Mathematik>
Subtraktion
Fluid
Automatische Handlungsplanung
Transportproblem
Transinformation
Systemplattform
Framework <Informatik>
Graph
Datensatz
Knotenmenge
Informationsmodellierung
Arbeitsplatzcomputer
Zusammenhängender Graph
Abstand
Softwareentwickler
Speicher <Informatik>
Maßerweiterung
Protokoll <Datenverarbeitungssystem>
Graph
Open Source
Einfache Genauigkeit
Routing
Plug in
Schlussregel
Vektorraum
Automatische Handlungsplanung
Modul
Objekt <Kategorie>
Mapping <Computergraphik>
Flächeninhalt
Last
Mereologie
Speicherabzug
Partikelsystem
Stapelverarbeitung
Bit
Prozess <Physik>
Komponententest
Punkt
Inferenz <Künstliche Intelligenz>
Formale Sprache
Bridge <Kommunikationstechnik>
Komplex <Algebra>
Richtung
Netzwerktopologie
Client
Unimodale Verteilung
Algorithmus
Einheit <Mathematik>
Mixed Reality
Skript <Programm>
Lineares Funktional
Shape <Informatik>
Sichtenkonzept
Datennetz
Systemaufruf
Profil <Aerodynamik>
Ein-Ausgabe
Teilbarkeit
Konfiguration <Informatik>
Arithmetisches Mittel
Menge
Dateiformat
Information
Instantiierung
Aggregatzustand
Zeichenkette
Standardabweichung
Rückkopplung
Subtraktion
Mathematisierung
Automatische Handlungsplanung
Term
Transportproblem
Knotenmenge
Bildschirmmaske
Logarithmus
Reelle Zahl
Datentyp
Zusammenhängender Graph
Softwareentwickler
Speicher <Informatik>
Touchscreen
Einfach zusammenhängender Raum
Diskretes System
Schlussregel
Einhüllende
Automatische Handlungsplanung
Generizität
Spezialrechner
Flächeninhalt
Last
Mereologie
Speicherabzug
Stapelverarbeitung
Resultante
Subtraktion
Punkt
Datensichtgerät
Mathematisierung
Güte der Anpassung
Profil <Aerodynamik>
Abfrage
Eliminationsverfahren
Automatische Handlungsplanung
Software Development Kit
Konfiguration <Informatik>
Chipkarte
Mapping <Computergraphik>
Unimodale Verteilung
Arbeitsplatzcomputer
Kantenfärbung
Resultante
Unimodale Verteilung
Abfrage
Garbentheorie
Automatische Handlungsplanung
Transportproblem
Resultante
Addition
Graph
Hochdruck
Abfrage
Automatische Handlungsplanung
Quick-Sort
Konfiguration <Informatik>
Videokonferenz
Bildschirmmaske
Unimodale Verteilung
Einheit <Mathematik>
Logarithmus
Algorithmus
Mereologie
Ablaufverfolgung
Lesen <Datenverarbeitung>
Softwaretest
Resultante
Konstruktor <Informatik>
Zentrische Streckung
Extrempunkt
Minimierung
Natürliche Zahl
Schlussregel
Automatische Handlungsplanung
Transportproblem
Dialekt
Objekt <Kategorie>
Unimodale Verteilung
Algorithmus
Festspeicher
Arbeitsplatzcomputer
Mereologie
Projektive Ebene
Portscanner
Instantiierung
Aggregatzustand

Metadaten

Formale Metadaten

Titel Tempus: A framework for multimodal trip planning
Alternativer Titel Geospatial - Tempus
Serientitel FOSDEM 2015
Autor Mercier, Hugo
Lizenz CC-Namensnennung 2.0 Belgien:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen 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.
DOI 10.5446/34397
Herausgeber FOSDEM VZW
Erscheinungsjahr 2016
Sprache Englisch
Produktionsjahr 2015

Inhaltliche Metadaten

Fachgebiet Informatik

Ähnliche Filme

Loading...
Feedback