NextGenIntelligentTransportaon: Measuring’People ... · • In partnership with SFMTA • 6 weeks • 800+ participants • Entry, exit survey • Location-aware survey app •

NextGen  Intelligent  Transporta2on:  Measuring  People,  Controlling  Things  

Raja  Sengupta,  Professor  Systems,  CEE,  UC  Berkeley  sengupta@ce.berkeley.edu,  

5107170632  

Berkeley  Terraswarm  project  

•  CENTER  MISSION:  The  TerraSwarm  Research  Center  (TSRC)  aims  to  enable  the  simple,  reliable,  and  secure  deployment  of  advanced  distributed  sense-­‐control-­‐actuate  applicaSons  on  shared,  massively  distributed,  heterogeneous,  and  mostly  uncoordinated  swarm  plaTorms  through  an  open  and  universal  systems  architecture.  

•  In  normal  operaSon,  a  swarm-­‐enabled  \Smart  City"  not  only  helps  run  the  infrastructure  more  eecSvely  but  empowers  its  occupants  by  providing  more  eecSve  interfaces,  beWer  mobility,  and  experiences  in  immersive  realiSes  in  a  way  not  possible  before.  For  example,  maintenance  crews  may  recruit  sensors  from  underground  uSliSes,  and  combine  that  sensor  data  with  data  from  pipe-­‐crawling  robots  and  from  the  cloud.  They  can  use  this  informaSon  to  guide  maintenance  operaSons  using  overlay  displays  in  a  manner  similar  to  what  televised  sporSng  events  use,  based  on  contextual  3D  informaSon.  

Human  and  Social  Capital  

•  Smart  CiSes  ,  Caragliu  etal  2009  – Urban  performance  currently  depends  not  only  on  the  city's  endowment  of  hard  infrastructure  ('physical  capital'),  but  also,  and  increasingly  so,  on  the  availability  and  quality  of  knowledge  communicaSon  and  social  infrastructure  ('intellectual  capital  and  social  capital').  ……  It  is  against  this  background  that  the  concept  of  the  smart  city  has  been  introduced  as  a  strategic  device  to  encompass  modern  urban  producSon  factors  in  a  common  framework  and  to  highlight  the  growing  importance  of  InformaSon  and  CommunicaSon  Technologies  (ICTs),  social  and  environmental  capital  ……  

The  benefits  of  a  city  investment  are  determined  by  the  produc2on  func2on  of  the  investment  and  the  use  its  

ci2zens  make  of  it.    

•  In  transportaSon  ciSzens  control  – AcSvity  choice  – Trip  chaining  –   Trip  choice  – Mode  choice  – Departure  Sme  choice  

– Route  choice  

The  inter-­‐play  of  the  physical  ar2fact    and  its  use  can  be  important  

•  In  transportaSon  ciSzens  control  – AcSvity  choice  –  Trip  chaining  –   Trip  choice  – Mode  choice  – Departure  Sme  choice  

–  Route  choice  •  From  

hWp://thestandinginvitaSon.wordpress.com/tag/game-­‐theory/      

The  User  Equilibrium  is  not    System  Op2mal  

•  In  transportaSon  ciSzens  control  –  AcSvity  choice  –  Trip  chaining  –   Trip  choice  – Mode  choice  –  Departure  Sme  choice  –  Route  choice  

•  Making  the  inter-­‐play  of  the  two  important  

•  From  hWp://thestandinginvitaSon.wordpress.com/tag/game-­‐theory/      

It’s  2002.  You  are  the  mayor  of  Seoul.  There  is  a  major  traffic  crisis  in  the  city,  with  congesSon  rising  by  as  much  as  5%  yearly.  You  have  £200  million  to  spend  on  solving  the  problem.  What  do  you  do?  Surely  you  build  more  roads  to  ease  the  traffic…  right?  Surprisingly,  the  mayor  did  the  opposite:  he  spent  the  money  demolishing  roads.  Even  more  surprisingly,  it  worked.  Because  of  the  strange  fact  of  Braess’s  Paradox,  shuing  down  roads  can  actually  decrease  traffic.  Here’s  how.  

To  realize  the  benefits  of  an  investment  one  must  Adapt  the  City  to  the  Ci2zen  and  The  Ci2zen  to  the  City  

Feedback  Control  Theory  is  one  science  of  Adapta2on  

ITS  is  older  on  the  Supply-­‐side  More  recent  on  Demand  

•  For  safety:  Collision  Warning,  Lane  Keeping,…  •  For  Increasing  Supply  

–  Ramp  Metering  –  Coordinated  Arterial  Signal  Control  –  Bus  fleet  management  systems  –  Electronic  Toll  CollecSon  

•  For  Managing  Demand  (CiSzen  Choices)  –  Cordon  Pricing/CongesSon  Pricing  – Dynamic  Parking  Pricing  –  Spare  the  Air  Day,  Bike  to  Work  Day  –  Travel  Feedback  Programs  

ITS  is  older  on  the  Supply-­‐side  More  recent  on  Demand  

•  For  safety:  Collision  Warning,  Lane  Keeping,…  •  For  increasing  supply  

–  Ramp  Metering  –  Coordinated  Arterial  Signal  Control  –  Bus  fleet  management  systems  –  Electronic  Toll  CollecSon  

•  For  Managing  Demand/User  Choice  –  Cordon  Pricing/CongesSon  Pricing  – Dynamic  Parking  Pricing  –  Spare  the  Air  Day,  Bike  to  Work  Day  –  Travel  Feedback  Programs  

Done  as  Feedback    control  

These  are  NOT  

The  difference  between  traffic  management  and  demand  management  is  the  difference  between  control  and  planning  

Survey  Op2mize  Intervene  

Deploy  

Parking  price  Cordon  price  Spare  the  air  day  

Link  speed  measurement  

Link  inflow  (op2mal)  control  

2me  

2me  

Ramp  Metering  

Demand  Management  

Evaluate  Baseline  

ITS  is  older  on  the  Supply-­‐side  More  recent  on  Demand  

•  For  safety:  Collision  Warning,  Lane  Keeping,…  •  For  increasing  supply  

–  Ramp  Metering  –  Coordinated  Arterial  Signal  Control  –  Bus  fleet  management  systems  –  Electronic  Toll  CollecSon  

•  Managing  Demand/AdapSng  User  Choice    –  AcSvity  choice  –  Trip  chaining  –   Trip  choice  – Mode  choice  –  Departure  Sme  choice  –  Route  choice  

Done  as  Feedback    control  

NOT  done  as  feedback  control    Survey  à  OpSmize  à  Intervene  à  Evaluate  

Why?  

The SF Travel Quality Study •  In partnership with SFMTA •  6 weeks •  800+ participants •  Entry, exit survey •  Location-aware survey app

•  Take Muni, use our app How do you feel? How was the ride? –  13,000 responses

•  Real-time subject location for 6 weeks –  82 million location points

•  Real-time transit vehicle location

Mobile survey •  How did this Muni

experience make you feel?

–  Pleased –  Frustrated –  Relaxed –  Impatient

•  How do you generally feel today?

•  How satisfied were you with…

–  Reliability –  In-vehicle travel

time –  Wait time –  Transfer time –  Accuracy of

real-time information

•  How satisfied were you with…

–  Cleanliness –  Crowding –  Pleasantness of

other passengers

–  Safety

EmoSons   OperaSons   Environment  

Matching the data Smartphone  locaSon  data  

Transit  vehicle  raw  AVL  data  

Transit  runs  

Personal  reliability  metrics  Route  shape  

files  

Personal  transit  

travel  diary  

Timetable  data  

SaSsfacSon  surveys  

2  

1  

3  

4  

•  Strong effect of in-vehicle delays on satisfaction and choice behavior

How do surveys relate to engineering?

•  Responses differ between online surveys and mobile surveys (Question presentation? Environment?)

Survey: Presentation, Context?

Daily  mobile  survey,  graphical  labels  L  -­‐  J  

End  of  study  mobile  survey,  graphical  labels  L  -­‐  J    

End  of  study  online  survey,  text  (Likert)  labels  

Quan2fied  Traveler  Part  1:  Automated  Travel  Diaries  

ICTBR  2012,    Jariyasunant  Phd  thesis  UCB    

Phone  App:  GPS,  Accelerometer,  WiFi,  Cell  Tower  data  +  Cloud  Server:  Processing  it,  serving  web  pages  

The  Automated  Trip  Diary  

•  No  User  Input  –  Automated  Signal  Processing  and  Machine  Learning  for  Trip  origin,  

des2na2on,  2me,  route,  mode  •  Learns  HotSpots  and  HotRoutes  •  Learns  HotModes  with  user  input  

Quantified Traveler

An Evaluation •  3 weeks March 18 to April 7, 2012 •  Pre and Post Experiment Survey

–  55 statements on awareness, experiential and instrumental attitudes, moral obligations, perceived norms, perceived control, self efficacy

•  Installed the QT App on personal Iphones and Android phones

•  Week 1 no feedback, 7th day given link to QT, Week 2 Spring Break data, Week 3 feedback –  Week 2 data not used in behavior change analysis

An Evaluation •  135 subjects

–  No one uninstalled the App, 118 completed both surveys –  4143 trips logged on 5 modes (bike, walk, drive, bus, train) –  258 km per subject on average –  Looked at website 4.1 times on average in week 2 –  Participants corrected mode on 13.5% of trips

•  78 subjects used for behavior change analysis

Users stop using the average applications quickly. Long term audiences are generally 1% of total downloads" - Pinch Media http://www.techcrunch.com/2009/02/19/pinch-media-data-shows-the-average-shelf-life-of-an-iphone-app-is-less-than-30-days

An Evaluation •  135 subjects

–  No one uninstalled the App, 118 completed both surveys –  4143 trips logged on 5 modes (bike, walk, drive, bus, train) –  258 km per subject on average –  Looked at website 4.1 times on average in week 2 –  Participants corrected mode on 13.5% of trips

•  78 subjects used for behavior change analysis

•  Basic Finding: Behavior change from week 1 to 2 –  Decrease in driving: Frequent drivers 38% reduction

(p<0.01), Infrequent drivers 27% (p=0.09) –  Increase in walking (p-value 0.03)

Travel Feedback Programs: Travelers meet with a counselor and receive personalized advice

High rates of success switching travelers to more sustainable modes

QT  as  a  con2nuous  feedback  System  

Interven2on  

Deploy  

2me  

Evaluate  Baseline  

Deploy  

Sense  

QT  is  a  con2nuous  feedback  system  

•  Its  design  theory  is  in  Psychology.  Not  a  control  theory.  

Daily  travel  

SmartPhone  Travel  Diary  

QT  Web  Server  

Time  Cost  Exercise  Green    

Traveler  Choices  

People  persuade  people.    Can  computa2onal  systems  persuade  people?  

Personalized  feedback  

Personalized  feedback  ?  

Successful Actuation needs Aesthetics and good HCI

•  Summary page liked the most followed by breakdown and trip diary. Liked comparison to others in the group. –  4 questions on 7-point Likert scale about the website.

•  I enjoyed taking a look at my dashboard/statistics/trip history page and getting a summary of my travel –  Mean = 5.4, Std Dev = 1.2

•  In the future, this web page is something I would consider using. –  Mean = 5.9, Std Dev = 1.0

•  If I were to set a goal to change my travel behavior (be greener, reduce cost, travel less), I consider this web page helpful. –  Mean = 5.1, Std Dev = 1.3

•  This web page was easy to use. Mean = 5.3, Std Dev = 1.1

Joint  work  with  

•  Andre  Carrel,  Venky  Ekambaram,  DJ  Gaker,  Dr.  Jerry  Jariyasunant,  Thejo  Kote,  Eric  Mai  

•  Several  undergraduate  students  •  Prof.  Joan  Walker,  Prof  Maya  Abou  Zeid  •  Spin-­‐off  Ventures  2010  –    

– automaSc.com  – Lockx.com  – Baytripper.org  – Human  Intellect  Lab  (hWp://myne.net/)  

Some  ITS  Theories  and  Systems  

•  For  safety:  Collision  Warning,  Lane  Keeping,…  •  For  increasing  supply  

–  Ramp  Metering  –  Coordinated  Arterial  Signal  Control  –  Bus  fleet  management  systems  –  Electronic  Toll  CollecSon  

•  Managing  Demand/AdapSng  User  Choice    –  AcSvity  choice  –  Trip  chaining  –   Trip  choice  – Mode  choice  –  Departure  Sme  choice  –  Route  choice  

Done  as  Feedback    control  

This  is  NOT  

Why?  

Because  the  technological  arSfacts  for  the  sensing  and  actuaSon  of  user  choice  did  not  exist  unSl  recently    On  the  supply-­‐side  they  have  existed  for  almost  two  decades  

Rise of Ubiquitous Computing: Continuous Sensing and Actuation for User Choice

•  Quantified Traveler is a part of the Quantified Self genre

-­‐  Record  Finances  -­‐  Set  Goals  -­‐  Analyze  Personal  Spending  -­‐  Analyze  Trends  

Behavior Change or Persuasive Technology

•  Quantified Self + •  Closing the loop with an objective

– With the consent of the subject –  Better eating, Better exercise, Sustainable travel –  Ubifit,  Fish  N  Steps,  Healthy  Lifestyle  Coach,  Designing  Games  to  Mo2vate  Physical  Ac2vity,  Persuasive  Picture  Frames  for  Proper  Posture,    

–  Persuasive  Conference  

Sensing  and  ActuaSon  Has  Arrived  for  Behavioral  Systems  

•  For  CiSzen  Choices  – Cordon  Pricing/CongesSon  Pricing  – Dynamic  Parking  Pricing  – Spare  the  Air  Day,  Bike  to  Work  Day  – Travel  Feedback  Programs  

Sensing  smartPhone  Fitbit  OFX  NEST  …..    Con2nuously,  Longitudinal    depth  

Sensing  and  ActuaSon  Has  Arrived  for  Behavioral  Systems  

•  For  CiSzen  Choices  – Cordon  Pricing/CongesSon  Pricing  

– Dynamic  Parking  Pricing  – Spare  the  Air  Day,  Bike  to  Work  Day  

– Travel  Feedback  Programs  

And  ActuaSon  Apps  Tweets  Facebook  Recommenda2on    engines….    Con2nuously,  Longitudinal    depth  

Pozdnoukhov  and  Coffey  

Bikeshare  Commute  Data  

TwiWer  Topic  Analysis  

Granger  Causality  Analysis  

Behavioral  Systems  as  Dynamical  Systems  What  part  is  generalizable?  

The  Berkeley  Approach  PlaTorms,  Colleagues,  ExploraSons  

Psychology,  Economics,  and  ComputaSon  

Shachar  Kariv,  Economics  

Scob  Moura,    Control,  Energy  

Alexei  Pozdnoukhov,  GIS  Computa2onal  Sociology  

Elizabeth  Deakin,  City  Planning  

Joan  Walker  Global  Metropolitan  Studies  

John  Canny  HCI,  Behavioral  Data  Mining  

Raja  Sengupta  Control  Theory  

Design  Theory  in  Psychology    QT  rests  on  the  Theory  of  Planned  Behavior  (Azjen  1991)  

•  Psychological  variables:  –  Awareness  of  travel  impacts  is  associated  with  more  favorable  

aitudes  towards  sustainable  transportaSon  behavior  –  Peer  influence  effect  is  significant:  

•  Favorable  comparison  of  travel  impacts  to  one’s  peers  leads  to  more  posiSve  aitudes  

•  Unfavorable  comparison  has  the  opposite  effect  

•  Behavior  change:  –  More  posiSve  aitudes  and  norms  are  associated  with  a  significant  

reducSon  in  driving  –  Regular  drivers  decrease  their  driving  most,  while  students  decrease  

walking/biking  distance  –  More  frequent  feedback  (logins  to  the  website)  is  associated  with  an  

increase  in  walking/biking  distance  

Change  in  walk  /bike  distance

Change  in  adtude

Change  in  awareness

Change  in  norms

Change  in  distance  driven

Peer  influence

Driver Student

Logins

Indicators

Indicators

Indicators

The  QT  Structural  EquaSon  Model  

Change  in  walk  /bike  distance

Change  in  adtude

Change  in  awareness

Change  in  norms

Change  in  distance  driven

Peer  influence

Driver Student

Logins

Indicators

Indicators

Indicators

The  QT  Structural  EquaSon  Model  

0.102  (0.66)

0.180  (1.35)

0.0352  (2.51)

-­‐0.611  (-­‐2.70)

0.00755  (0.43)

0.0117  (0.93)

-­‐0.224  (-­‐2.16)

0.0125  (2.01)

-­‐0.121  (-­‐3.94)

-­‐0.737  (-­‐3.35)

0.0433  (2.99)

GeneralizaSon  is  done  by  demographics:  Age,  Income,  …  

For  PersonalizaSon:  There  is  an  incompleteness  problem  

Separate  into  Es2ma2on  and  Control  Experimental  Micro-­‐Economics  à  Func2onal  Forms  for  the  Es2ma2on  

Problem  •  Four  fundamental  tradeoffs  

–  Risk  vs.  Return  –  Today  vs.  Tomorrow  – Work  vs.  Leisure  –  Self  vs.  Others  

•  ComputaSonal  toolkits  to  reveal  parameters  (preferences)  -­‐  Recoverability  

$100  

$40  $30  

$30  

Risk  vs.  Return  How  do  you  choose?  

Heads  

Tails  

Kariv  2007  

EsSmate  parameters  for  Gul’s  model  loss  and  disappointment  aversion    With  CARA  specificaSon    

The  Laboratory  is  in  the  Wild  

•  Four  fundamental  tradeoffs  –  Risk  vs.  Return  –  Today  vs.  Tomorrow  – Work  vs.  Leisure  –  Self  vs.  Others  

•  ComputaSonal  toolkits  to  reveal  preferences  

$100  

$40  $30  

$30  

Risk  vs.  Return  How  do  you  choose?  

Heads  

Tails  

Kariv  2007  

EsSmate  parameters  for  Gul’s  model  loss  and  disappointment  aversion    With  CARA  specificaSon    

Android  Screen  

Time  of  Day  Effect   Day  of  Week  Effect  Effect  of  Sme  of  day  on  percentage  allocated  to  the  cheapest  asset.  ID  73261:                ID  74188:  

Effect  of  day  of  the  week  on  percentage  allocated  to  the  cheapest  asset.  ID  77848:                ID  62896:  

The  dynamics  of  preference?  Insufficient  data  

Sensing  and  ActuaSon  Has  Arrived  for  Behavioral  Systems  

•  For  CiSzen  Choices  – Cordon  Pricing/CongesSon  Pricing  – Dynamic  Parking  Pricing  – Spare  the  Air  Day,  Bike  to  Work  Day  – Travel  Feedback  Programs  

Sensing  smartPhone  Fitbit  OFX  NEST  …..    Con2nuously,  Longitudinal    depth  

Sensing  and  ActuaSon  Has  Arrived  for  Behavioral  Systems  

•  For  CiSzen  Choices  – Cordon  Pricing/CongesSon  Pricing  

– Dynamic  Parking  Pricing  – Spare  the  Air  Day,  Bike  to  Work  Day  

– Travel  Feedback  Programs  

And  ActuaSon  Apps  Tweets  Facebook  Recommenda2on    engines….    Con2nuously,  Longitudinal    depth