Cassandra Workshop Presentation

1. © 2009 IBM Corporation
An IT view of Smarter Cities
Jurij Paraszczak for Smarter Cities Global Team
Director Industry Solutions and Smarter Cities
IBM Research
jurij@us.ibm.com
With many thanks to the Research Smarter Cities team

2. © 2008 IBM Corporation
The city – a system of systems
Systems from transportation to energy, healthcare, commerce, education,
security, food, water, jobs and economic growth come together and interact
with each other
How can they be managed better ?
2 15 September 2010
EDUCATION • TRANSPORTATION • SOCIAL SERVICES • SAFETY • UTILITIES • HEALTHCARE • COMMUNCATION
EDUCATION • TRANSPORTATION • SOCIAL SERVICES • SAFETY • UTILITIES • HEALTHCARE • COMMUNCATION
+
$

3. © 2008 IBM Corporation
Overview
Smarter Cities approach creates solutions
which simplify the way in which the myriad
city operations act in a city and helps city
managers make rational decisions based on
data and prediction
Over 100 + people are working around the
world are learning with our customers and
deploying models and analytics which use a
common platforms and approaches to enable
repeatable processes
From this work we are discovering patterns
and approaches which help in this
simplification, reducing cost and providing
new insights
Taking advantage of our deep scientific and
engineering capabilities in IBM Research
Asset
Management
Pipes, Roads,
Wires, Bldgs,
etc.
Resource
Optimization
Water, traffic,
energy etc.
People
Motivation &
Inclination
System of
Systems
Safety & Security
City water, energy,
buildings & transport
Jobs
Comfort
Lifestyle
City Needs

4. © 2008 IBM Corporation
Rio
Emerg.
Natural
Resources
Texas
River Basin
NY Bldgs,
Emer Security
Ranaana
Water
IBM Research: Smarter City Global engagements
Smarter City
Activity
Dublin Traffic,
Water, Energy
Shenyang
Water, CarbonBeijing
Traffic
Delhi
Energy
Traffic
Traffic Agency
West Coast
PA Bldgs
PNW
SmartGrid
Singapore
Traffic Water
Tokyo
Integ. City
Bornholm
Energy
Dubuque
Water,
Energy
DC WASA
Water
Beijing
Energy
Sydney
Energy
Melbourne
Energy &
LifeScience
Moscow
Nanotech
Stockholm
Traffic

5. © 2009 IBM Corporation
Analysing Cities
Who wants what when and where

6. © 2008 IBM Corporation
Who spends what in cities ?
IBM assessment from top 50 cities by
population
3 City types identified
Mature Large
Mature Medium
Cities in Transition
Each city type has different focus
Mature Large - safety & security
Mature Medium - maintenance and resource
management
In Transition - focus on new state of art
infrastructure and resource management
systems
Mature Large
$285B
(19 Cities/198M People)
Mature Medium
$115B
(16 Cities/59M People)
Cities In Transition
$161B
(15 Cities /217 M People)
50 Cities Budget : $561B
City Budgets in Aggregate

7. © 2008 IBM Corporation
IBM Smarter Cities Challenge
The Smarter Cities Challenge is a competitive grant program awarding $50
million worth of technology and services over the next 3 years to 100 cities
around the globe. These grants are designed to address the wide range of
financial and infrastructure challenges facing cities today
See http://smartercitieschallenge.org/

8. © 2008 IBM Corporation
Observations in working around the world with Cities
Key issues include
Ability to engage with citizens and engage their opinions and support
Management of public safety
Scheduling of work and activities in the face of conflicting or completely
non integrated activity. Dig patch Dig
Understanding of movement of people and traffic in city
Caused by
Lack of understanding of details of what is happening in city
And use of data and analytics to determine same

9. © 2008 IBM Corporation
We are targeting the following city domains
Traffic &
Transportation
Building Energy
Water
availability &
purity
Safety

10. © 2009 IBM Corporation
Underlying Science and Engineering
From paper to models

11. © 2008 IBM Corporation
Developing the Research which underlies Smarter Cities
We view the Smarter City through this structure
Infrastructure Technologies & Tools
Data Models Optimization
Business
Decisions
Emerging area: Human interaction with Smarter City
Core Technologies
Solutions

12. © 2008 IBM Corporation
Understanding disconnects: A warning and a simple example of a
common problem

13. © 2008 IBM Corporation
Using mathematics and models to drive the business activity - for
example, traffic management
Operational/
Transactional
Road Usage
Optimization,
GHG emission
models
•More granular
charging, by location
•Analysis of traffic
patterns to manage
city congestion.
•Modeling traffic to
predict and manage
entire system
•Dynamic and
congestion based
pricing
•Route planning and
advice, shippers,
concrete haulers,
limo companies,
theatres, taxis etc
•City-wide, dynamic
traffic optimization
•Charge collection
only - disconnected
operational data
•Transaction data from
the management of
payments
•Little automated use
is made of real-time
traffic data
Business
Development
Operational/ Transactional Insights System wide control
2008-10 2008-12? 2009-15?

14. © 2008 IBM Corporation
Advanced Analytics
is the use of data and models to provide insight to guide decisions
Models
Analytics
Data
Insight
Data sources:
Business automation
Instrumentation
Sensors
Web 2.0
Expert knowledge
“real world physics”
Model:
a mathematical or
algorithmic
representation of
reality intended to
explain or predict
some aspect of it
Decision executed
automatically or
by people

15. © 2008 IBM Corporation
Stockholm
Traffic
Managing Traffic in Stockholm

16. © 2008 IBM Corporation
Stockholm Road Charging
40 Gantries with 18 ingress
points
Approx 320K entries/exists
per day

17. © 2008 IBM Corporation
Charging to reduce traffic

18. © 2008 IBM Corporation
Case Study – Stockholm Congestion Charging
Main objective – to reduce congestion by
between 10% and 15%.
Project – to build a system that would
automatically tax Swedish registered
vehicles entering and leaving the city centre
between 6.30 and 18.30, Monday to Friday
(excluding national holidays).
Duration – 7 months (January - July 2006)
Challenges – political sensitivity, public
scrutiny, referendum at the end of the trial
to decide on whether to implement the
congestion tax permanently
Results
Traffic congestion in Stockholm was reduced
by 25%, far above the original target
Traffic queuing times fell by up to 50%.
Journey times were faster and more
predictable
Stockholm bus timetables were re-written to
take improvements to traffic flow into
account
Pollution levels in the city fell by between
10% and 15%
Confidence in the system was high due to
minimal enforcement and administrative
errors
Scheme was re-launched in August 2007 after the
public referendum voted in favour of the system

19. © 2008 IBM Corporation
Analysing Traffic

20. © 2008 IBM Corporation
20
Time to Action
Notional Information Supply Chain for Decision-making
Transforming the Information Supply Chain to reduce the time to action!
SOURCES
Elapsed Time to Action
WAREHOUSE
Reports
Ad-hoc Queries
DATA INTEGRATION
OPERATIONAL DATA STORES
DATAMARTS
Bus Process
& Event Mgmt
Operational
Reports
Dashboards Planning Scorecarding
Analytical Modeling & Information
Stream computing represents a critical paradigm shift
Analytical Modeling
& Information

21. © 2008 IBM Corporation
Infosphere Streams in Stockholm - why models are important
Slow/stop Moderate Average Good
Fast >140
Km/hr
Bouillet, Riabov, Verscheure
Traffic Speed

22. © 2008 IBM Corporation
Predicting Traffic

23. © 2008 IBM Corporation
Traffic Prediction Tool (TPT) – background and motivation
The ability to capture the current traffic state and to project it to the near future from
available data sources is critical for real-time traffic management
Traditional data sources Non-traditional data sources
Inductive loop Traffic camera
Infrared laser radar Passive infrared – ultrasonic sensor
GPS device Smart phone
Historical origin-destination trip tables
Fixed locations,
sparse in the network
…

24. © 2008 IBM Corporation
Traffic Prediction Tool (TPT)
Model: stochastic model used to predict traffic in Singapore
Little automated use is made of the gigabytes of real-time
traffic data today; often, by the time it is received, it is no
longer representative of the actual traffic
► Issue: “real-time” is too late
IBM’s TPT provides a layer of intelligence by using sensor
data in sophisticated algorithms that create relevant
insights from the raw data
► IBM Innovation: forecast the future
TPT accurately
forecasts future
traffic conditions,
including incidents
0 50 100 150 200 250
1000200030004000
rr
r
rrr
r
r
r
r
r
r
r
rr
r
r
r
r
rr
r
r
rr
r
r
r
r
r
volume
blue = forecast black = actual red = incident
time
Current Focus
Traffic Operations:
Variable Message Sign
setting; traffic signal
timing, ramp metering
Future Use
Traffic Planning; Dynamic
Road Pricing; congestion
based tariff setting; route
planning & advice
Extension: Data Expansion
(2008 IME) develop algorithm to fill in
gaps of real-time sensor data, resulting
in a complete picture of future traffic
state, network-wide
tool screenshot
results

25. © 2008 IBM Corporation
Agent Based Analytics and prediction

26. © 2008 IBM Corporation
Large-scale Agent-based Traffic Flow Simulator
IBM Mega Traffic Simulator
output
Traffic
census
Map data
Driving log
traffic volume for each link
base data
IBM Mega Traffic Simulator
Origin-
destination
Road
network
Driver Model
input
3k cars/hour
0.5k cars/hour
2k cars/hour
Link A Link C
CO2emission
Link B
CO2 emission for each link
Traffic situation with more than the millions of vehicles can be simulated.
Traffic flow with various types of drivers behavior model can be simulated.
Traffic situation with more than the millions of vehicles can be simulated.
Traffic flow with various types of drivers behavior model can be simulated.
IBM Zonal Agent-based Simulation Environment
Agent Space
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Agent
Simulation Space
Messaging
Handler
Communication Manager
Thread Manager
threadthreadthreadthreadthreadthread
Memory Manager
Agent Manager
Message Queue
Scheduler
Java Virtual Machine
Driver
Agent
Vehicle
Driver Behavior Model

27. © 2008 IBM Corporation
Application of the simulator: What-If Analysis
The simulator provides an experimental environment for traffic policy makers to perform what-if analysis
concerning traffic in a large city.
Current traffic status
If Condition1 Then …………
If Condition2 Then …………
If Condition3 Then …………
If Condition4 Then …………
How the traffic policy and city-
design should be in the aging
society?
What is the appropriate
information providing service to
minimize traffic congestion?
How the traffic would change if
we introduce congestion tax.
What is the proper traffic policy to
solve traffic congestion, green issues....
How the total emission would
change if we introduce a new
traffic policy?
32k cars/day32k cars/day32k cars/day32k cars/day
49k cars/day49k cars/day49k cars/day49k cars/day
2k cars/day2k cars/day2k cars/day2k cars/day

28. © 2008 IBM Corporation
Water Infrastructure Management
DC WASA
Water

29. © 2008 IBM Corporation
Analytics Driven Asset Management (ADAM)
Data Operational, Failure, Usage, Condition, Customer, Location
DataAssets
•Asset Management
•Work Management
•Service Management
•Inventory / Contract
•Procurement Management
EAM/SCADA
Enterprise Asset Management Scada, Sensors, Inspection, Metering Systems
ADAM
Descriptive, Predictive and
Prescriptive Analytics
Insight,
Foresight and Prescriptions
•Maintenance Planning
•Maintenance Scheduling
•Replacement Planning
•Condition Assessment
•Failure Cause Analysis
•Failure Prediction
•Usage Analysis
•Customer Analysis

30. © 2010 International Business Machines Corporation 30
ADAM: Analytics Driven Asset Management
Predictive analytics models enabling “fix before
break”
Spatial Schedule Optimization enables “while in
the neighborhood “ scheduling
Data analytics enable forecasting of water usage
and detection of usage anomalies
130,000Water Meters
1,600,000Sewer Customers
600,000Water Customers
370MGallons / dayWaster Water Capacity
36000Catch Basins
24,000Valves
9000Hydrants
1800 MilesSewer Pipes
1200 MilesWater Pipes
All from conventional historical and log data!

31. © 2008 IBM Corporation
Optimization
GIS Data
Predictive Analytics
ADAM for Water Utilities V1.0
EAM
Advanced Reporting
Spatio-Temporal
Manual Scheduling
Failure Pattern and
Cause Analysis
Customer
Segmentation
Automated spatial
schedules
Failure Risk based
PM Optimization
Usage & Revenue
Forecasting
Automated Task level
rolling scheduling
Failure Prediction
Usage Anomaly
Detection
Dynamic Mobile
Work Management
Replacement
Planning
Non-Revenue Water,
Energy Optimization
Work
Management
Predictive
Maintenance
Usage/ Revenue
Optimization
Water Usage Data

32. © 2008 IBM Corporation
Examples of Advanced Reporting – Catch Basin Work
Orders
Temporal Analysis of Work Order
Patterns
Spatial Distribution of annual work
Catch basic problem code
distributionWork classification vs Problem code visualization
Catch Basin

33. © 2008 IBM Corporation
Use cases
Task Level Scheduling
Dynamic Re-Scheduling using GPS data
ADAM V1.0 Use cases
• Manual Map Based Schedule Construction
• Semi-Automated Route Completion
• Multi-crew automated scheduling
Ongoing R & D

34. © 2008 IBM Corporation
IBM Research: Smarter City Global engagements
Smarter City
Activity
Dublin Traffic,
Water, Energy

35. © 2008 IBM Corporation
Smarter Cities Technology Centre
Dublin

36. © 2011 IBM Corporation
Developing technology to continuously assess the state of the public transport system and
provide personalized, real-time advice to riders and dynamic load-balancing opportunities
to transit providers
Background
– GPS & other sensor technologies are transforming
transportation analytics
Working closely with Dublin
– Demonstration visualisation of transportation
network status & guidance for bus drivers
Challenges
– Extracting insights from real-time, noisy, irregular
samples
– Taking actions under uncertainty with low latency
– Large volume & diversity of data
Transportation

37. © 2011 IBM Corporation
Dublin Bus – Demonstration

38. © 2011 IBM Corporation
Platform for gathering and analyzing Dublin city data,. Working with
Dublin City on an Open Innovation Platform for Cities
Background
– Governments are seeking to spawn & exploit
innovation & promote awareness through better
access to data of citizen’s interest
Deploying significant common infrastructure for
IBM’s SC community
– Common compute, data & network platform
– Data repositoru
– Connectivity into Dublin Systems
Challenges
– Data & model management in City-scale
environment
– Tools enabling domain experts to interface with
complex data & analytic challenges intuitively
City Fabric
Open
Collaborative
Research
Common
Standards &
Definitions
Advanced City Technology
Multi-City &
International
Collaboration
Platform
Data
Presentation
Open Innovation Platform

39. © 2008 IBM Corporation
Managing Public Safety in NYC and Chicago
NY City + Chicago
Public Safety

40. © 2008 IBM Corporation
Safety and Security Management
Chicago’s Virtual Shield Program
Implemented one of the most advanced city-wide intelligent security systems
The engagement is a part of Chicago's Operation Virtual Shield, a project that
encompasses one of the world's largest video security deployments
In the first phase, IBM helped the City experts and network engineers design and
implement a monitoring strategy infrastructure to capture, monitor and fully index
video for real-time and forensic-related safety applications
Korea Incheon Free Economic Zone
Implemented a public safety infrastructure with intelligent video monitoring as part
of the U-safety City project
Built a public safety system utilizing high-resolution cameras to view and monitor
activities to prevent crime and even predict possible events by recognizing and
analyzing certain patterns and data in real time

41. © 2008 IBM Corporation
Statistical modeling, machine learning & pattern recognition are key
technologies to enable Smart Safety and Security
Blob Tracking
Algorithm
Object
Classification
Algorithm
Color
Classification
Algorithm
Background
Subtraction
Algorithm
Machine learning enables recognition of person attributes
Statistical Modeling is the key to handling change

42. © 2008 IBM Corporation
Selected Research & Technical Challenges
Handling crowded scenes
Finer grained analysis of objects
Federated / Partitioned Architectures
Analytics at the edge

43. © 2008 IBM Corporation
Managing Energy in Buildings
NY Bldgs,

44. © 2008 IBM Corporation
i-BEE (IBM Building Energy and Emission) Analytics ToolSet
Saving energy, improving energy efficiency and reducing greenhouse gas (GHG)
emissions are key initiatives in many cities and municipalities and for building owners
and operators.
For example, New York City's government spends over $1 billion a year on energy, and is
committed to reducing the City government's energy consumption and CO2 emissions by 30% by
2030 (PlaNYC). Buildings emit about 78 percent of the city’s GHG emissions. NYC plans to
invest, each year, an amount equal to 10% of its energy expenses in energy-saving measures.
In order to reduce energy consumption in buildings, one needs to understand patterns
of energy usage and heat transfer as well as characteristics of building structures,
operations and occupant behaviors that influence energy consumption.
i-BEE is physics, statistics and mathematics based building energy analytics that
Assess how different energies are used (and GHG is emitted) in different ways
Benchmark energy (GHG emission) uses among peer buildings
Track energy consumption and its changes due the improvement actions (e.g., retrofits)
Forecast future energy consumption (and GHG emission)
Simulate impacts of various changes (improvements) on energy consumption and GHG emission
Optimize energy consumption, efficiency and GHG emission

45. © 2008 IBM Corporation
Modeling Approach

46. © 2008 IBM Corporation
Dashboard – Example (Energy Use & Greenhouse Summary, GIS
Energy Intensity Map)
K-12 Schools

47. © 2008 IBM Corporation
The Benefit of Analytics
Identify anomaly that can lead to failure of equipment and wasted energy, and
take corrective actions for faults
Statistical Analysis (SPC, CUSUM, Time Series Model, Data Mining..)
Identify underperforming buildings with respect to peer buildings and identify
the root causes
Multiple Regression Modeling
Accurately estimate heat loss (gain) through walls, roofs, windows, and develop
retrofit plans
Heat Transfer Model
Identify key characteristics of building structures, operations and behaviors that
influence energy consumption and take actions for modifications
Forecast future energy consumption and develop cost effective procurement
plan of energy
Forecasting Model
And others…

48. © 2009 IBM Corporation
The Role of People in Cities
Dubuque

49. © 2008 IBM Corporation
IBM Research: Smarter City Global engagements
Dubuque
Water,
Energy

50. © 2008 IBM Corporation
Green Dubuque
CICERO: Citizen centric Intelligence & Resource Optimization

51. © 2008 IBM Corporation
Participants Compete – IBM provides the platform
Pilot defined
Each week, individual households and teams will have the chance to win prizes.
Each week, you will be randomly assigned to a team made up of 3-5 other Pilot
members.
You will not know your other team members but you can chat with them using the
team chat on the site.
Each week, individual households and teams will win prizes and/or will be
registered to win our mid-way and final prizes!
Prize drawings take place at the end of week 6 and at the end of week 1
IBM provides
Cloud platform and software that aggregates and maps usage
Provides metrics and competition information
Tracks all usage helping development of behavioural models

52. © 2008 IBM Corporation
CICERO deployed for Resource Consumption Management
Cloud-based real-time intelligence & interaction for instrumented, interconnected
cities
•Deployed for water silo and work underway for electric silo
•Resource optimization & decision support for maximizing city performance
•Models & Incentives for changing citizen resource consumption behavior
•Interest from multiple cities to join cloud delivered service

53. © 2008 IBM Corporation
Whither Weather

54. © 2008 IBM Corporation
The opportunity and challenge of combining models
Weather models and resulting damage prediction for Electric
Utilities
IBM Weather Prediction System DEEP THUNDER - accurate to 2 km
x 2 km area
A mathematical model that describes the physics of the atmosphere
– The sun adds energy, gases rise from the surface, convection causes
winds
Numerical weather prediction is done by solving the equations of these
models on a 4-dimensional grid (latitude, longitude, altitude, time)
Solution yields predictions of surface and upper air
– Temperature, humidity, moisture
– Wind speed and direction
– Cloud cover and visibility
– Precipitation type and intensity
Challenge is to predict business impact of weather

55. © 2008 IBM Corporation
IBM uses advanced weather forecasting technologies to predict power
demand and outages - Deep Thunder our unique world class weather
prediction technologies
Weather causes damage and outages
Outages require restoration (resources)
Restoration takes time, people, etc.
Build stochastic model from weather observations, storm damage and
related data
Outage location, timing and response
Wind, rain, lightning and duration
Demographics of effected area
Ancillary environmental conditions
Weather
prediction
Power Line
Damage
prediction
Restoration
time
prediction
Work crew
requirement
prediction

56. © 2008 IBM Corporation
13 March 2010 Nor’easter Deep Thunder Impact Forecast
Actual Outages (Repair Jobs) Estimated Outages (Repair Jobs)

57. © 2008 IBM Corporation
Approach to Urban Flood Forecasting
Precipitation
Estimates
Flood
Prediction
Impact
Estimates
Model
Calibration
Refine Sensor Network
and Model CalibrationActual Flood Impacts
Weather Prediction and/or
Rainfall Measurements
Analysis of Precipitation

58. © 2008 IBM Corporation
Integrating Systems

59. © 2008 IBM Corporation
IBM Research: Smarter City Global engagements
Rio
Emergency
Management

60. © 2008 IBM Corporation
RIO Operations Center
Allows diverse agencies to share
emergency information and plan
coordinated responses
Part of Rio's preparatory efforts
for Brazil's hosting of soccer's
World Cup in 2014 and the city's
hosting of the 2016 Olympic
Games.
Components include
Data acquisition and integration
center from multiple agencies
High Resolution Weather
Prediction System coupled to
hydrological flooding models
Traffic management systems
Emergency operations
Integrated scheduling,
optimization and allocation of
processes

61. © 2008 IBM Corporation
Summary
IBM Research is focusing our global resources on the understanding and
management of resource usage and deriving an understanding of how these
resources interact
The integration of technology, mathematics. IT and computer science
coupled with advances in algorithms, processor speed communication
bandwidth are enabling the management of cities in ways previously
unimaginable
World pressures from emissions, population and economic growth are driving
ever increasing efficiency in the use of every resource
The Smarter Cities approach enables this transition