SlideShare a Scribd company logo
TomTom Dynamic Routing Technology
Heiko Schilling
Navigate to …                    … how it works




       On-board            Off-board: routes.tomtom.com


             Identical Software Stack
             both on-board & off-board
             Both on-board & off-board
Navigate to …   … how it works
Navigate to …   … how it works
Navigate to …   … how it works
Navigate to …   … how it works
Navigate to …   … how it works
Navigate to …                           … how it works



     Did you know that:
     -  Up to 15 alternative routes
        between START & FINISH
     -  But People use/know only 1 or 2
       [Jansen & Den Adel, 1987] [Bovy & Stern 1990]

     -  Unawareness causes traffic jams
Standard Speed Input: Flat Speed Profile (Single Value)

A8 Zaanstad ➔ AMS




Confidential & Internal - TomTom International b.v.
                                                      9
10
IQR Speed Input: Time-Dependent Speed Profile

A8 Zaanstad ➔ AMS, towards town




Confidential & Internal - TomTom International b.v.
                                                      11
IQR Speed Input: Time-Dependent Speed Profile

AMS ➔ A8 Zaanstad, outwards town




Confidential & Internal - TomTom International b.v.   69
                                                      12
IQR Speed Input: Step Function of Speed Values

AMS ➔ A8 Zaanstad, outwards town




per profile one look-up table with 7x24x12 values (5 min slots)

Confidential & Internal - TomTom International b.v.
                                                              13
Route Planning with Time-Dependent Speeds

                                      Destination @ 17:00




                   B
   A @ 15:59
                  12 min @ 15:55 - 15:59



        @ 15:00
Start
                                                       14
Route Planning with Time-Dependent Speeds

                                              Destination @ 17:02


                                              Destination @ 17:07

               C     13 min @ 15:55 - 15:59


                     13 min @ 16:00 - 16:05
        @ 16:00
        (+1 min) A     18 min @ 16:00 - 16:05

                                    12 min @ 15:55 - 15:59

Start
                                                               15
Route Planning with Time-Dependent Speeds

                                    17:02

                                     17:07




          16:00
          (+1 min)

15:00
        The network in front of you is changing with the minute
        ... and so does your fastest route!



                                                                  16
Time-expanded Network




                                                  Time




       Implicitly we are working on a much more
       complex time-expanded network.               17
Continuous Re-calculation

                             17:02

                               17:07




        16:00      continuously search in the back
        (+1 min)   for faster routes or ETA adjustment

15:00              •  How many bifurcation points?
                   •  How many route calculations?
                   •  Timeliness?
                   •  Better de-tours?
                   •  Different approach possible?
                                                         18
TomTom’s Challenge: To find the Best Route
amongst all Options for any Traffic Situation

Europe Map: 50 million crossings, 120 million roads….
          and 800 sextillion (21 zeros) possible routes
         All routes together consume space equal to 50 billion times the current size of the internet
                              Calculating all routes would take 25 trillion years




        So finding the best route takes time …
 but we’ve found a solution that is FAST & EXACT                                                          .
 TomTom IP: on- & off-board solution for FAST & EXACT Route Planning (Apollo)
 [WO 2011004026: NAVIGATION DEVICES AND METHODS CARRIED OUT THEREON],[WO 2011004029: NAVIGATION DEVICES]…




  Over 100 scientific publications in Mathematics/CS Community
  [Schilling, PhD’06], [Lauther, ‘04], [Sanders, Schulte, ESA’08], [Delling, ESA’09], [Goldberg, SODA’02] …
On-Board Planning Time (200 km route)
          TomTom   10   20   30   40      50      60 sec.
          100%
             -
             -
             -
             -
          95%
             -
             -
             -
             -
Quality




          90%
             -
             -
             -
             -
          85%
             -
             -
                                  Competitor Systems
             -
             -
          80%
Fast & Exact Routing: On-Board Planning Time (200 km)
      TomTom




    20% BETTER* routes in Europe
    30% BETTER* routes in North America
           (*) compared to standard routing technologies
FAST & EXACT Routing – Off-Board Planning Time: 10–35ms
  200.000 routes requests categorized according to distance:
      •  without Apollo we broke up the worker after a few days and 143.000 calculations
      •  with Apollo 200.000 OD pairs were calculated in a few hours

                 FAST & EXACT Routing Worker




          Standard Routing Worker
FASTFast EXACTRouting Demo
           & & Exact Routing Demo
                  DEMO
Standard Calculation Time: 35 Seconds      Fast Calculation Time: 1 Second




 2 identical devices
                              2 identical devices
     • ARM9 CPU, 266 MHz, 64 MB RAM, 32 bit memory bandwidth
          ARM9 CPU, 266 MHz, 64 MB RAM, 32 bit memory bandwidth
 IQ Route Request on a Benelux 2008.10 Map with a 9 Mb sidefile
    IQSchengen, Luxembourg Benelux 2008.10 Map with a NL (City
     •  Route Calculation on a (City Center) à Eemshaven, 9 Mb sidefile
     Center)
               Schengen, Luxembourg à Eemshaven, Netherlands
     • Identical results: 597 km in 5:43 hrs with 38 instructions
              Identical results: 597 km – 5:43 hrs – 38 instructions
FAST & EXACT Routing Demo




    Standard                                 FAST
Calculation Time:                     Calculation Time:
   35 seconds                             1 second




                      2 identical devices
      ARM9 CPU, 266 MHz, 64 MB RAM, 32 bit memory bandwidth

 IQ Route Calculation on a Benelux 2008.10 Map with a 9 Mb sidefile
           Schengen, Luxembourg à Eemshaven, Netherlands
         Identical results: 597 km – 5:43 hrs – 38 instructions
Apollo: Exact Route Planner

Luxembourg 2008.10 map
•  Identical        ETT w/o Apollo                 = 86%
•  Faster           ETT with Apollo                = 14%
•  Slower           ETT with Apollo                =     0%      IQR improved
                                                                 35% of routes
Benelux 2008.10 map
•  Identical        ETT w/o Apollo                 = 83%         Apollo improves
•  Faster           ETT with Apollo                = 17%
                                                                 20+ % on top
•  Slower           ETT with Apollo                =     0%


WCE & NAM 2008.10 map
•  Faster ETT with Apollo                          = 20 %
(non-Apollo standard A* route planner using all heuristics vs. Apollo
route planner; Route comparison on ~ 10.000 random origin-destination pairs)
Sunday Lunch Time: Apollo saved you 12 minutes of 1:26 hrs

       A* std route                Apollo route
Thursday afternoon: Apollo saved you 13 minutes of 1:11 hrs

       A* std route                 Apollo route
User Benefits of
FAST & EXACT Route Planning
User Benefit: Fast Re-Planning
User Benefit: Dynamic Routing
Continuously Searches for Better Route
User Benefit: Dynamic Routing
Continuously Searches for Better Route
User Benefit: Dynamic Routing
Continuously Searches for Better Route
Best Possible Route at Any Point in Time
Based on Historic Traffic/Speed Profiles
                                                                   London (W) à London (E)
Travel Time [minutes]




                        Monday   Tuesday   Wednesday   Thursday   Friday   Saturday   Sunday

                                               Time [hours in week]
User Benefit: Routing Time Machine
Preview Best Route in Advance
User Benefit: Routing Time Machine
Preview Best Route in Advance

More Related Content

PDF
EPOSA - real time positioning in austria
PPT
TcpGPS
PPT
Dynamic Routing All Algorithms, Working And Basics
PPT
Dynamic routing under_sec
PPT
Dynamic Routing RIP
PPT
Dynamic Routing
PDF
Allan Rasmussen TomTom Maps June 16th 2014 - INSPIRE Conference
PPSX
Scala @ TomTom
EPOSA - real time positioning in austria
TcpGPS
Dynamic Routing All Algorithms, Working And Basics
Dynamic routing under_sec
Dynamic Routing RIP
Dynamic Routing
Allan Rasmussen TomTom Maps June 16th 2014 - INSPIRE Conference
Scala @ TomTom

Viewers also liked (19)

PDF
TomTom Q1 2014 Financial Results
PDF
Dhcp commandas
PPT
Dynamic Routing IGRP
PPTX
PPTX
Sept 2017 dynamic routing
PDF
PPT
TomTom Presentation.
PPT
DOC
Mba Tomtom merger TeleAtlas
PPTX
Dynamic routing protocols (CCNA)
DOCX
Dhcp cisco
PPTX
Design and Implementation of Dynamic Routing in Wireless Networks
PDF
CCNAv5 - S2: Chapter 6 Static Routing
PDF
Class 7
PPT
Static Routing
PPTX
Secure Shell(ssh)
PPTX
Telnet & SSH Configuration
PPSX
Packet Tracer Tutorial # 2
TomTom Q1 2014 Financial Results
Dhcp commandas
Dynamic Routing IGRP
Sept 2017 dynamic routing
TomTom Presentation.
Mba Tomtom merger TeleAtlas
Dynamic routing protocols (CCNA)
Dhcp cisco
Design and Implementation of Dynamic Routing in Wireless Networks
CCNAv5 - S2: Chapter 6 Static Routing
Class 7
Static Routing
Secure Shell(ssh)
Telnet & SSH Configuration
Packet Tracer Tutorial # 2
Ad

Similar to TomTom Dynamic Routing (20)

PDF
FIWARE - Estimated Time Arrival Implementation Montevideo
PDF
Platforming_Automated_And_Quickly_Beamer
PPTX
A force directed approach for offline gps trajectory map
PDF
GPS for Interactive Traffic Analysis af Kristian Torp, AAU
PPTX
Global Map Matching using BLE Beacons for Indoor Route and Stay Estimation
PPT
L12.FA20.ppt
PDF
How Zalando accelerates warehouse operations with neural networks - Calvin Se...
PPTX
DIGITAL FARE METER
PDF
Geographic Computation in Perl
PPT
Darcy Bullock Presentation 2-3-10
PDF
MapReduce with Hadoop
PPTX
Eposa english
PDF
IoT at railways proposal
PPT
Feeder Separation Project
PDF
Lisa's Master's Thesis Project
PDF
ACM DEBS Grand Challenge: Continuous Analytics on Geospatial Data Streams wit...
PDF
Veridict Trafiklab meetup 2016 12-06
PPT
PDF
mnNOG 2020: The Journey [100% ROA Coverage]
PPT
TcpET
FIWARE - Estimated Time Arrival Implementation Montevideo
Platforming_Automated_And_Quickly_Beamer
A force directed approach for offline gps trajectory map
GPS for Interactive Traffic Analysis af Kristian Torp, AAU
Global Map Matching using BLE Beacons for Indoor Route and Stay Estimation
L12.FA20.ppt
How Zalando accelerates warehouse operations with neural networks - Calvin Se...
DIGITAL FARE METER
Geographic Computation in Perl
Darcy Bullock Presentation 2-3-10
MapReduce with Hadoop
Eposa english
IoT at railways proposal
Feeder Separation Project
Lisa's Master's Thesis Project
ACM DEBS Grand Challenge: Continuous Analytics on Geospatial Data Streams wit...
Veridict Trafiklab meetup 2016 12-06
mnNOG 2020: The Journey [100% ROA Coverage]
TcpET
Ad

Recently uploaded (20)

PDF
Agricultural_Statistics_at_a_Glance_2022_0.pdf
PPT
Teaching material agriculture food technology
PDF
Unlocking AI with Model Context Protocol (MCP)
PPTX
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
PDF
Approach and Philosophy of On baking technology
PDF
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
PDF
cuic standard and advanced reporting.pdf
PPTX
MYSQL Presentation for SQL database connectivity
PDF
Machine learning based COVID-19 study performance prediction
PDF
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
PDF
The Rise and Fall of 3GPP – Time for a Sabbatical?
PPTX
VMware vSphere Foundation How to Sell Presentation-Ver1.4-2-14-2024.pptx
PDF
Per capita expenditure prediction using model stacking based on satellite ima...
PDF
Diabetes mellitus diagnosis method based random forest with bat algorithm
PDF
KodekX | Application Modernization Development
PDF
Building Integrated photovoltaic BIPV_UPV.pdf
PDF
Encapsulation theory and applications.pdf
PDF
NewMind AI Weekly Chronicles - August'25 Week I
PDF
Encapsulation_ Review paper, used for researhc scholars
PPTX
Programs and apps: productivity, graphics, security and other tools
Agricultural_Statistics_at_a_Glance_2022_0.pdf
Teaching material agriculture food technology
Unlocking AI with Model Context Protocol (MCP)
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
Approach and Philosophy of On baking technology
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
cuic standard and advanced reporting.pdf
MYSQL Presentation for SQL database connectivity
Machine learning based COVID-19 study performance prediction
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
The Rise and Fall of 3GPP – Time for a Sabbatical?
VMware vSphere Foundation How to Sell Presentation-Ver1.4-2-14-2024.pptx
Per capita expenditure prediction using model stacking based on satellite ima...
Diabetes mellitus diagnosis method based random forest with bat algorithm
KodekX | Application Modernization Development
Building Integrated photovoltaic BIPV_UPV.pdf
Encapsulation theory and applications.pdf
NewMind AI Weekly Chronicles - August'25 Week I
Encapsulation_ Review paper, used for researhc scholars
Programs and apps: productivity, graphics, security and other tools

TomTom Dynamic Routing

  • 1. TomTom Dynamic Routing Technology Heiko Schilling
  • 2. Navigate to … … how it works On-board Off-board: routes.tomtom.com Identical Software Stack both on-board & off-board Both on-board & off-board
  • 3. Navigate to … … how it works
  • 4. Navigate to … … how it works
  • 5. Navigate to … … how it works
  • 6. Navigate to … … how it works
  • 7. Navigate to … … how it works
  • 8. Navigate to … … how it works Did you know that: -  Up to 15 alternative routes between START & FINISH -  But People use/know only 1 or 2 [Jansen & Den Adel, 1987] [Bovy & Stern 1990] -  Unawareness causes traffic jams
  • 9. Standard Speed Input: Flat Speed Profile (Single Value) A8 Zaanstad ➔ AMS Confidential & Internal - TomTom International b.v. 9
  • 10. 10
  • 11. IQR Speed Input: Time-Dependent Speed Profile A8 Zaanstad ➔ AMS, towards town Confidential & Internal - TomTom International b.v. 11
  • 12. IQR Speed Input: Time-Dependent Speed Profile AMS ➔ A8 Zaanstad, outwards town Confidential & Internal - TomTom International b.v. 69 12
  • 13. IQR Speed Input: Step Function of Speed Values AMS ➔ A8 Zaanstad, outwards town per profile one look-up table with 7x24x12 values (5 min slots) Confidential & Internal - TomTom International b.v. 13
  • 14. Route Planning with Time-Dependent Speeds Destination @ 17:00 B A @ 15:59 12 min @ 15:55 - 15:59 @ 15:00 Start 14
  • 15. Route Planning with Time-Dependent Speeds Destination @ 17:02 Destination @ 17:07 C 13 min @ 15:55 - 15:59 13 min @ 16:00 - 16:05 @ 16:00 (+1 min) A 18 min @ 16:00 - 16:05 12 min @ 15:55 - 15:59 Start 15
  • 16. Route Planning with Time-Dependent Speeds 17:02 17:07 16:00 (+1 min) 15:00 The network in front of you is changing with the minute ... and so does your fastest route! 16
  • 17. Time-expanded Network Time Implicitly we are working on a much more complex time-expanded network. 17
  • 18. Continuous Re-calculation 17:02 17:07 16:00 continuously search in the back (+1 min) for faster routes or ETA adjustment 15:00 •  How many bifurcation points? •  How many route calculations? •  Timeliness? •  Better de-tours? •  Different approach possible? 18
  • 19. TomTom’s Challenge: To find the Best Route amongst all Options for any Traffic Situation Europe Map: 50 million crossings, 120 million roads…. and 800 sextillion (21 zeros) possible routes All routes together consume space equal to 50 billion times the current size of the internet Calculating all routes would take 25 trillion years So finding the best route takes time … but we’ve found a solution that is FAST & EXACT . TomTom IP: on- & off-board solution for FAST & EXACT Route Planning (Apollo) [WO 2011004026: NAVIGATION DEVICES AND METHODS CARRIED OUT THEREON],[WO 2011004029: NAVIGATION DEVICES]… Over 100 scientific publications in Mathematics/CS Community [Schilling, PhD’06], [Lauther, ‘04], [Sanders, Schulte, ESA’08], [Delling, ESA’09], [Goldberg, SODA’02] …
  • 20. On-Board Planning Time (200 km route) TomTom 10 20 30 40 50 60 sec. 100% - - - - 95% - - - - Quality 90% - - - - 85% - - Competitor Systems - - 80%
  • 21. Fast & Exact Routing: On-Board Planning Time (200 km) TomTom 20% BETTER* routes in Europe 30% BETTER* routes in North America (*) compared to standard routing technologies
  • 22. FAST & EXACT Routing – Off-Board Planning Time: 10–35ms 200.000 routes requests categorized according to distance: •  without Apollo we broke up the worker after a few days and 143.000 calculations •  with Apollo 200.000 OD pairs were calculated in a few hours FAST & EXACT Routing Worker Standard Routing Worker
  • 23. FASTFast EXACTRouting Demo & & Exact Routing Demo DEMO Standard Calculation Time: 35 Seconds Fast Calculation Time: 1 Second 2 identical devices 2 identical devices • ARM9 CPU, 266 MHz, 64 MB RAM, 32 bit memory bandwidth ARM9 CPU, 266 MHz, 64 MB RAM, 32 bit memory bandwidth IQ Route Request on a Benelux 2008.10 Map with a 9 Mb sidefile IQSchengen, Luxembourg Benelux 2008.10 Map with a NL (City •  Route Calculation on a (City Center) à Eemshaven, 9 Mb sidefile Center) Schengen, Luxembourg à Eemshaven, Netherlands • Identical results: 597 km in 5:43 hrs with 38 instructions Identical results: 597 km – 5:43 hrs – 38 instructions
  • 24. FAST & EXACT Routing Demo Standard FAST Calculation Time: Calculation Time: 35 seconds 1 second 2 identical devices ARM9 CPU, 266 MHz, 64 MB RAM, 32 bit memory bandwidth IQ Route Calculation on a Benelux 2008.10 Map with a 9 Mb sidefile Schengen, Luxembourg à Eemshaven, Netherlands Identical results: 597 km – 5:43 hrs – 38 instructions
  • 25. Apollo: Exact Route Planner Luxembourg 2008.10 map •  Identical ETT w/o Apollo = 86% •  Faster ETT with Apollo = 14% •  Slower ETT with Apollo = 0% IQR improved 35% of routes Benelux 2008.10 map •  Identical ETT w/o Apollo = 83% Apollo improves •  Faster ETT with Apollo = 17% 20+ % on top •  Slower ETT with Apollo = 0% WCE & NAM 2008.10 map •  Faster ETT with Apollo = 20 % (non-Apollo standard A* route planner using all heuristics vs. Apollo route planner; Route comparison on ~ 10.000 random origin-destination pairs)
  • 26. Sunday Lunch Time: Apollo saved you 12 minutes of 1:26 hrs A* std route Apollo route
  • 27. Thursday afternoon: Apollo saved you 13 minutes of 1:11 hrs A* std route Apollo route
  • 28. User Benefits of FAST & EXACT Route Planning
  • 29. User Benefit: Fast Re-Planning
  • 30. User Benefit: Dynamic Routing Continuously Searches for Better Route
  • 31. User Benefit: Dynamic Routing Continuously Searches for Better Route
  • 32. User Benefit: Dynamic Routing Continuously Searches for Better Route
  • 33. Best Possible Route at Any Point in Time Based on Historic Traffic/Speed Profiles London (W) à London (E) Travel Time [minutes] Monday Tuesday Wednesday Thursday Friday Saturday Sunday Time [hours in week]
  • 34. User Benefit: Routing Time Machine Preview Best Route in Advance
  • 35. User Benefit: Routing Time Machine Preview Best Route in Advance