Relating different (traffic) data in the NDW Observatory at DiTTlab - - PowerPoint PPT Presentation

“From struggling through data towards searching in information”

Relating different (traffic) data in the NDW Observatory at DiTTlab Edoardo Felici (NDW)

NDW: a unique alliance of 19 authorities

Traffic information

Up-to-date, complete and unambiguous

Traffic management

Central source for all road authorities

Traffic policy and traffic research

Goals

• Less traffic jams
• Safer roads
• Less emissions

Happy road users

DiTTlab scope

Virtual Reality, Gaming

Evaluation & assessment

(Real-time) diagnostics, estimation & prediction

Mixed Reality

(BIG) Data Processing

(Open-source) Multiscale Simulation

Overall DiTTlab architecture

• An integrated framework
• (Open) data from all

possible sources:

– Traffic – Transport – Networks

• Advanced data

assimiliation and analyses

• Open source multi-scale,

multi-modal traffic and transport simulations E - OpenTrafficSim Ontology F - GIS

(semi-static data: transport infra & built environment) C - Database (dynamic data: traffic, transport, weather, etc)

G - GUI’s / Editors A - OpenTraffic Simulator H - Visualisers, analysers, exporters

D - data import

B - OpenTrafficSim Input & toolset

(Calibratie, Validatie, Identificatie, Fusie, Assimilatie tools)

11

Mild day

June 8 2015

Heavy rainfall

May 31 2015 April 28 2015

Roadside fire

March 24 2015

Large accident

NDW project: an intelligent database

• From struggling through data to searching through information

A20, June 9 2015

Large accident

March 24 2015

Large accident Large accident

A16, January 9 2015

Java / Javascript / leaflet

• Webtool to join

and present different data types

Dittlab.tudelft DB:

• 12447 Cross sections
• 66589 measurement points

(5,1M considering changes over time)

• ALL Geo functionality

Netezza DB:

• 2,382,489,886 lane/usrclass

records dynamic measurements (March 2015)

• ca 1/10 of that roadway

records dynamic data Cross-section

Detectors

• multiple

measurement points

GetDailyWeath er GetASMStats GetCrossections

GetMeaspoints

GetDailyEvents

GetWeather

GetEvents

GetRWTotals

GetDailyTotals

NDW-tool

Different congestion patterns with different causes

15

16

CLASS 1 CLASS 2 CLASS 3

• Step 1 – train the database to recognize congestions patterns
• Date and time
• Route characteristics

(geo & digraph info)

• weather (rainy, 10o

Celsius, etc)

• Incidents / events
• Etc.

(Available) Metadata Clustered congestion graphs

Pictures:

• Don’t have to be the same size
• Don’t need the same amount of pixels
• Don’t need to have the same length-width

ratio

• Do need to have the same colormap
• Do need the same time-space scale (to

detect shockwaves)

• Type X
• Type Z
• Type Y
• Travel time loss
• Travel time spread
• Data quality
• % freight
• Demand patterns
• Etc.

from databases / GIS: To define/calculate

Super fast searches in a mega database

• Step 2 – Dissect the graphs in a “bag of features”

(just like we do when recognizing faces)

Feature descriptions Clustering of features SURF algorithm K-means algorithm Extract keypoints

Super fast searches in a mega database

• Step 3 – Every traffic pattern can be summarized in a histogram of features

(a vector with a number)

Traffic pattern Feature vectors Class 1 Class 2 Class 3

Super fast searches in a mega database

Unknown pattern Classification Training model Dissection in “features” Pattern type X

• Step 4 – Classification of new patterns

So far so good … TO-DO: iterative process (1) Start classification with limited data set (2) Manually refine classification (3) Retrain SVM classifier (4) Back to (1)

Super fast searches in a mega database

• Prototype of Congestion-Search-engIne (CoSi)

– Iterative refinement of the SVM classifier (until classification is accurate and precise enough) – Develop smart routines to find and classify all congestions patterns – Perform the classification – SVM + available metadata = Search-index

• Refine and expand with weather- and other metadata
• Build WebGUI for CoSi

This year:

Super fast searches in a mega database

That could go better!

Lab measurements

Just like CSI: data fusion gives more, better and more robust evidence (1+1=3) Examples of extra context:

• Circumstances (CAN

data, weather, incidents, events, news, etc)

• Network state &

alternatives (ITS, parking, Public transport data)

• OD, route- and modal

split (GSM, Apps, public transport)

• Driving experience, skill

and style (Apps, CAN data)

• Travel motives (social

media, apps)

Field measurements

DiTTlab next steps: how do we get data?

Real behaviour?

• 1. Many more (GIS-, mathematic-, modelling-)

toolsets and skills needed

• 2. Working in interdisciplinary teams

➡Traffic engineers ➡Data scientists ➡Computer scientists

• 3. Sharing data, sharing expertise (NL is too

small!)

DiTTlab next steps: a different way of working!

Just like CSI: data fusion gives more, better and more robust evidence (1+1=3) Examples of extra context:

• Circumstances (CAN

data, weather, incidents, events, news, etc)

• Network state &

alternatives (ITS, parking, Public transport data)

• OD, route- and modal

split (GSM, Apps, public transport)

• Driving experience, skill

and style (Apps, CAN data)

• Travel motives (social

media, apps)

• Based on NDW data and many other potential data

sources (like bicycle data)

• Estimation of main unknowns:
• Volumes, vehicle loss hours
• Inflows and turns (what comes on and what goes off)
• Capacities, critical speeds
• Network fundamental diagramme
• Estimation of origin-destination information on

different scale levels

• Simulate (predict?) traffic based on available

information on different scale levels

• Multi-scale estimation of all relevant variables

Plans 2016 and beyond

Dynamics 1-10 seconds Using CAM-bus, volumes

• n lanes, travel times,

local weather etc.

Multi-scale estimation of variables

Dynamics 1 minute – few minutes Using travel times, volumes, travel times,

• rigin-destinations local

weather etc. Dynamics 5-30 minutes Using CAM-bus, volumes

• n lanes, travel times,
• rigin-destinations, modal

split, regional weather etc.

Hans van Lint Edoardo Felici j.w.c.vanlint@tudelft.nl edoardo.felici@ndw.nu