Quoi de Neuf?* Leveraging Things and People as Sensors for Smart CitiesAnimesh PathakProject-team MiMoveInria Paris-RocquencourtBIS 2014

June 19, 2014

* ’ssup?

The Internet of Things (IoT)

• Connecting the physical world with the virtual world• Leveraging sensing and actuation capabilities embedded in

Things

• Increasing number of mobile Things with many

benefits => mobile IoT

www.gizmag.com www.robotshop.comwww.slashphone.com

- 2

- 3

Mobile IoT: Illustrating scenario

Wind-chill in Paris?

How to measure wind-chill?

How to know they are in Paris?

Wind-chill sensor?

Location sensor?

? ?

! So many devices! Diverse types! Unknown availability! Incomplete measurements

Service Oriented

Middleware

http://www.netatmo.com/

! Coarse grained data

Focus: - On-demand- Discrete- Sensing/actuating requests

- 4

Key Challenges of the Mobile IoT

• Heterogeneity (Representation)• Different Things, different data types

• Ultra Large Numbers• Millions of Things, large volumes of data

• Dynamic Environment • Mobile Things, short-life span

More likely to not register

Enough coverage

Design Rationale- Do not register redundant

Things to reduce number- Set coverage threshold

Managing Scale: Probabilistic Registration

Centralized Approach• Compute decision at

registry as search problem• Using global displacement knowledge• But computation time

increases linearly

5

Decentralized Approach• Compute decision on Thing• Estimate displacements

of registered nodesusing mobility models

• Constant computation timeJardin des Tuileries

✔

6

Managing Scale: Probabilistic Look-up

Candidate ThingSelected Thing

Probabilistic Look-up: Normal distribution

7

r3

Dmax

Lookup controls participation of Things

Xr1

r2

r4

Candidate ThingSelected Thing

Moving on… to moving in the city

- 8

Photo: syracuse.com

Transport: in Europe

- 9

Photo: Animesh Pathak

Transport: This is what “Smart” means now

- 10

Photo: ParisByTrain

Photo: Anjali Sharan

Transport: This is what “Smart” means now

- 11

Photo: ParisByTrain Photo: SNCF/Android Photo: DMRC/AppStore

Photo: Anjali Sharan

Same info, smaller screen!

… but there are more things to do!

- 12

Our Inspiration: Connected Things, Smart Cities, Mobile Citizens

- 13

http://revistamoviles.blogspot.com.es/2012/03/telefonica-convertira-alcala-de-henares.html

Scenario

14

A football match just ended and the train station is crowded. Charlie is going to Alice’s place with his son and he does not want to catch the Metro 7 if it is crowded.

Photo: ReutersPhoto: Dailymail

Idea: Participatory Sensing, for Mass Transit

- 15

Like Waze, but for the Metro, Bus, RER…

TravelDashboard: Participatory Sensing for Urban Transport

- 16

- 16

EIT ICT Labs• Ambientic• Inria• Thales• ALU• UCL• Systematichttps://www.rocq.inria.fr/arles/traveldashboard

Prototype: TravelDashboard Paris

- 17

Prototype: TravelDashboard Paris (2)

00 MOIS 2011 - 18

Meanwhile in London… new incentives!

- 19

and now… a short video…

- 20

https://www.youtube.com/watch?v=mZyWWwm-dN8

Back to Paris… and to industrial adoption!

- 21

from Passenger to Operator…

- 22

and now… another short video…

- 23

http://vimeo.com/69621153

3cixty Platform (2014-)Platform for comprehensive view for city travelers

• Uses Traveldashboard capability.

• Provides preference based inputs to the user by querying related data.

• Customizable platform

24

- 25

Incentives• Financial• Ego-centric• Altruistic• Democratic

Physical and Social KB

Participatory Sensing Social Sensing

So

cial

-sem

anti

cs-

bas

ed a

cces

s-

con

tro

l p

oli

cies

Per

son

ali

zati

on

CitizensOpportunistic Sensing

Needed: Algorithms and protocols for incentive and privacy

Recognition and Future Directions

• Awarded the “Prix Mashup” (€2500) at the RATP OpenDataLab in

May 2013• Led to appearance on TVFIL78 evening programming

• Federated Social Networking• Privacy policies for mobile social networks

• Incentive Mechanisms• How to encourage people to share their inputs

• Project “Sarathi”, and Inria Lab CityLab (http://citylab.inria.fr)• Tackling smart-city challenges in Europe, India, and the Silicon Valley!

- 26

MERCI!

Members of MiMove@Inria (previously known as ARLES)

Especially Sara Hachem

The TravelDashboard and 3cixty project partners

Animesh.Pathak@inria.fr