workshop on cyber-physical systems platforms – tânia calçada “urbansense platform”

32
UrbanSense Platform Porto Living Lab Tânia Calçada [email protected] Center of Competence for Future Cities of the University of Porto Instituto de Telecomunicações

Upload: future-cities-project

Post on 07-Aug-2015

55 views

Category:

Technology


1 download

TRANSCRIPT

UrbanSense Platform Porto Living Lab

Tânia Calçada [email protected] Center of Competence for Future Cities of the University of Porto Instituto de Telecomunicações

Future Cities project goals

Expand Centre of Competence in Future Cities of U. of Porto to

Promote Inter-Disciplinary Research and

Transfer Knowledge to Industry in Northern Region of Portugal

5/18/2015 2

1: Form inter-disciplinary teams

2. Build testbeds for urban-scale experiments

3. Work closely with end users from day one

4. Share data sets

5. Work with Industry Partners

7

Companies

Institutional Support

6. Bring the results out to the world

Cloud

Porto Living Lab

Porto Living Lab: On the ground

Crowdsensing application SenseMyCity

SenseMyCity: Smart phone application designed to apply to different research projects

• Gather data from choice of available sensors – GPS, magnetometer, accelerometer – Wi-Fi and Bluetooth devices

• Supports external sensing devices – OBD to collect vehicle’s data – Heart wave and heart beat sensor

• Secure transmission and storage

• What to do with data? – Fuel consumption estimation – Mobility patterns – Route similarity (e.g. car sharing) – WiFi coverage maps

Vehicular Ad-hoc Network

Vehicles connect to each other – V2V

to the infrastructure - V2I

• Largest vehicle-to-vehicle (V2V) testbed in the world – 600 vehicles

• Two independent test beds – BusNet: buses and garbage trucks in the city of Porto – HarborNet: trucks and boats in Leixões harbor

• Joint work with – Universidade do Porto – Universidade de Aveiro – VeniamWorks – Instituto das Telecomunicações

5/18/2015 12

Bus Net: free WiFi on board

• Started Sep 2014

• 150k

– Unique users

• 14.4 TB

– Internet traffic

• 1.3M

– Internet sessions

UrbanSense platform

Large-scale infrastructure for local monitoring

Data Collecting Units

Environmental sensors

Pedestrian counters

Low cost devices

Cover a restricted

area

Cloud data services

Storage: database

Processing: calibration

Analysis: data fusion

Sharing: open data

5/18/2015 14

UrbanSense goals and aplications

Understand and get aware of environmental and behaviour phenomena

Impact in the City

• City operations • Identify critical urban areas

• Detect events automatically

• Evaluate impact of urban interventions

• Companies • Test products

• Validate business models

Research

• Open data

• Big data, data mining

• Wireless networks

• Cyber physical systems

• Urban planning

• Transportation

• Climate

• Environment

• Health

Applications

• Pollution early warnings

• Waste collection management

• Garden smart management

• Smart parking

• Localization

• Surveillance

• Real estate

5/18/2015 15

UrbanSense multi-disciplinar ongoing research

• Health – asthma and air air pollution – Morbidity vs cold spells or heat waves

• Traffic and urban planning – Act in traffic policies to reduce noise and/or air quality – Solar radiation vs coatings on roads and facades – Smart artistic lighting

• Design and social sciences – Feed data to social meeting places – Sensor enclosure integration in urban environment

5/18/2015 16

UrbanSense platform architecture

• Data collection Units (DCUs) – Sensors

• Pedestrian counters • Environmental sensors

– Mobile and static DCUs locations • 50 in the roof of buses • 25 environmental in static locations • 60 pedestrian counters in static locations

• Data communications: WiFi – Large availability around the city – City hotspots: Porto Digital or Eduroam – Mobile hotspots offered by vehicular network – Use buses as data mules

• Data storage, processing, analysis and sharing – Relational database – Open data: API to access data

• Based on REST • Based on Fi-Ware

5/18/2015 17

Data Collecting Units

• SW and HW developed within the project • Processing board: Raspberry Pi

– Local data analysis and storage – Manage intermittent communications

• Conditioning circuit electronics – Control Board

• Custom made expansion board for Rpi

– Sensor board • Host embedded sensors but exposed to elements

• Low cost sensors • 2 Wi-Fi interfaces (1st phase static DCUs)

– City hotspots: Management (and data upload) – Opportunistic communications: Data upload

• Enclosure and shield – Costume made

5/18/2015 18

Noise Air quality , RH,

temperature Processing, storage

and control

Solar Radiation

sensor

WiFi interfaces

Weather Station

UrbanSense cloud architecture

5/18/2015 19

Opportunistic communications: data mules

Low cost communications to collect data that is tolerant so some delay

5/18/2015 20

Data Collecting

Unit Road Side Unit

Data Collecting

Unit

CloudData Base

Fib

er

Op

tic

WiFi WiFi

WAVE

Proof of concept

Developed in the context of a Master thesis

5/18/2015 21

Sense Unit Road Side Unit

On-Board Unit

Sensors configuration overview

DCUs T1: 15 units

DCUs T2: 10 units

DCUs T3: 50 units

Counters 60 units

520 sensors

Air Quality

Particles ✓ ✓ ✓ 75

Carbon monoxide (CO) ✓ ✓ ✓ 50

Ozone (O3) ✓ ✓ ✓ 75

Nitrogen dioxide (NO2) ✓ ✓ ✓ 75

Meteoro logical

Temperature & Humidity (RH) ✓ ✓ ✓ 75

Luminosity ✓ ✓ ✓ 75

Anemometer, pluviometer, wind vane ✓ 10

Solar radiation ✓ 10

Noise ✓ ✓ 25

Counters (based on video camera) ✓ 50

5/18/2015 23

Calibration and Validation Methodology

Compare measurements of UrbanSense sensors with reference sensors

• Use reference sensors – Expensive and homologated devices

– Typically used by environmental scientists

– Provided by U. Porto research groups

• Collect UrbanSense and reference data – Same location and time

• Create a Machine learning model – Inputs:

• Gas sensor measurements

• Temperature and humidity

– Output: • Gas concentration

5/18/2015 24

Provided by LEPABE

Data Mining model

Sensors Location

Traffic lights Balconies

5/18/2015 25

Picture by EPFL

Static deployment location plan

• Covered zones – Industrial

– Park

– Traffic

– Touristic

– Waterside

• 25 static DCUs – 2 DCUs

installed in Summer 2014

– 23 DCUs planned to Spring 2015

5/18/2015 26

1ST DCU AT R. FLORES ON THE 22TH JULY 2014 2ND DCU AT R. DAMIÃO DE GOIS ON THE 12TH AUGUST 2014 3RD DCU AT FEUP ON THE 23TH APRIL 2015

UrbanSense Platform Deployment

Sensors in a Flowerpot at R. Flores, Porto

5/18/2015 28

The Sensors

A – Wind direction B – Wind speed C – Precipitation

D – Temperature and humidity

E – Noise

F – Solar radiation

C A D

E F

B E

F

5/18/2015 29

Sensors in a Traffic light pole

5/18/2015 30

Sensors in a video surveillance pole

Workshop on cyber-physical systems

• 09h30 – Future Cities Project - UrbanSense Platform Tânia Calçada, FEUP/IT

• 10h00 – Sigfox network in Portugal Pedro Costa, NarrowNet

• 10h30 – COFFEE BREAK AND NETWORKING

• 11h00 – Citibrain

Rui Costa, Ubiwhere Rui Rebelo, Micro I/O

• 12h00 – Smart Water Metering Customer Consumption Francisco Cardoso, Águas do Porto

5/18/2015 32