SlideShare ist ein Scribd-Unternehmen logo
1 von 38
Downloaden Sie, um offline zu lesen
The 2010 Simulated Car Racing
Championship @ GECCO-2010
Daniele Loiacono, Luigi Cardamone,
Martin V. Butz, and Pier Luca Lanzi
2




               2010 Simulated Car Racing Championship
                     9 races during 3 conferences
                       ACM GECCO-2010, Portland, OR (USA), July 7-
                      IEEE WCCI-2010, Barcelona (Spain), July 18-23
                   IEEE CIG-2010, Copenhagen (Denmark), August 18–21


                        Develop a driver for TORCS
                     (hand-coded, learned, evolved, …)

            Drivers will be awarded based on their score
                   in each conference competition

          At the end, the team with highest overall score
                      wins the championship


2010 Simulated Car Racing Championship @ GECCO-2010
3




                        What is the structure of a race?

           Three stages: warm up, qualifiers, actual race

            During warm-up, each driver can explore the
                  track and learn something useful

        During qualifiers, each driver races alone against
         the clock (the best 8 drivers move to the race)

             During the race all the drivers race together



2010 Simulated Car Racing Championship @ GECCO-2010
Motivations


      Proposing a relevant game-based competition
          more representative of commercial games AI
          more similar to a real-world problem
      Proposing a funny and exciting competition
          you can see and play with the entries of this competition
          human players can interact with AI
          a lot of programmed AI available for comparison
      Proposing a challenging competition
          not designed with Machine Learning in mind
          computationally expensive
          real-time
          dealing with a lot of practical issues




2010 Simulated Car Racing Championship @ GECCO-2010
What‘s new?


     If everything seems under control, you're not going fast enough
                                                      — Mario Andretti
      Warm-up stage
          Before qualifying stage, competitors have 100000 game
          ticks to race on the track
          Allows track learning and optimization of parameters
      Noisy sensors
          Track sensors and opponent sensors are affected by a
          Gaussian noise (standard deviation equal to 10% of the
          readings)
      Extended sensor model
          Focus sensors
          Z position and speed
          Direction of track sensors fully customizable
          Added clutch control and focus command



2010 Simulated Car Racing Championship @ GECCO-2010
The Open Racing
Car Simulator
The Open Racing Car Simulator


      TORCS is a state of the art open source simulator written in C++

      Main features
          Sophisticated dynamics
          Provided with several
          cars, tracks, and
          controllers
          Active community of
          users and developers
          Easy to develop your
          own controller




      OS Support
          Linux: binaries and building from sources
          Windows: binaries and ―limited‖ bulding from sources support
          OSX: legacy binaries and no building from sources support 




2010 Simulated Car Racing Championship @ GECCO-2010
The Open Racing Car Simulator
     & the Competition Software




            TORCS                                           TORCS
                                                        PATCH

       BOT      BOT      BOT                            SBOT    SBOT   SBOT


     The competition server
                                                      UDP       UDP      UDP
        Separates the bots from TORCS
        Build a well-defined sensor model
        Works in real-time
                                                      BOT       BOT      BOT



2010 Simulated Car Racing Championship @ GECCO-2010
Sensors and actuators


      Rangefinders for edges on the track and opponents
      Speed, RPM, fuel, damage, angle with track, distance race,
       position on track, etc.




      Four effectors: steering wheel [-1,+1], gas pedal [0, +1],
       Brake pedal [0,+1], Gearbox {-1,0,1,2,3,4,5,6}
2010 Simulated Car Racing Championship @ GECCO-2010
Competitors
The competitors


      Four entries submitted to this first leg
          AUTOPIA, Madrid and Granada
          J. Muñoz, Carlos III University of Madrid
          S.Pohl, J. Quadflieg and T. Delbrügger, TU Dortmund
          Joseph Alton, University of Birmingham

      Two more entries from the 2009 championship
         COBOSTAR
         (T. Lönneker & M.V. Butz, University of Würzburg)
         POLIMI (Cardamone, Politecnico di Milano)




2010 Simulated Car Racing Championship @ GECCO-2010
AUTOPIA




  Industrial Computer Science Department.
  Centro de Automática y Robótica
  Consejo Superior de Investigaciones Científicas
  Madrid, Spain
  Contact:E. Onieva (enrique.onieva@car.upm-csic.es)
2010 Simulated Car Racing Championship @ GECCO-2010
Architecture Schema


      Three basic modules for gear, steering and speed control

      Steady stage genetic algorithm to compute the best weights
       to combine parameters for steering and target speed control

      Opponents module
         Acts on steering and brake signal to overtake opponents
         and avoid collisions

      Learning Module in Warm-up Stage
          Factors over the target speed in certain track segments




2010 Simulated Car Racing Championship @ GECCO-2010
Opponents Management (Steering)


      If O(-20,-10,10,20) detects an opponent with lateral
       displacement less than 20 meters, move the steering
       trackwidth/100 through the direction with more free
       distance.
      If O(-40,-30,-20,-10) detects an opponent with lateral
       displacement less than 20 meters, move the steering
       trackwidth/50 through the right.
      If O(-70,-60,-50)<15 meters, move the steering
       trackwidth/50 through the right.
      If O(-140,-130,-120,-110,-100,-90,-80)<15meters , move
       the steering trackwidth/50 through the right.
      Mirroring actions for positive sensors

      O(X)  Opponent Measure of Sensor Oriented at X

      Steering Modification limited to 0.35
2010 Simulated Car Racing Championship @ GECCO-2010
Learning Module (Warm-Up)


      Running normally in warm-up stage.

      Maintain a vector with as many real values as the track
       length in meters.

      Vector initialized to 1.0

      If the vehicle goes out of the track or suffers damage then
       multiply vector positions from 100 meters before the current
       position by 0.9.




2010 Simulated Car Racing Championship @ GECCO-2010
Mr Racer




  Susanna Pohl, Jan Quadflieg and Tim Delbrügger
  TU Dortmund


2010 Simulated Car Racing Championship @ GECCO-2010
Mr. Racer 2009-2010


      Mr Racer 2009
          Good classifier which identifies six situations
          Acceleration/brake learned offline using an EA
          Model of the track learned online
          Simple heuristic to use the model: override the learned
          behaviour on straights and in full speed corners to drive
          flat out
      Mr Racer 2010
          Save the model after warmup, use it during
          qualifying and the race
          Use the model to derive a plan consisting of
          target speeds and a racing line
          Optimize the parameter set of the planning module,
          left to be done for the next round


2010 Simulated Car Racing Championship @ GECCO-2010
Mr. Racer - Classification

      Angle based measure mapped to six situations (straight, fast
        corner, slow corner, etc)




2010 Simulated Car Racing Championship @ GECCO-2010
Mr. Racer 2010 – The catch


      Noise completely breaks our classifier

      Without a descent classifier we can‗t learn the track

      Without a trackmodel we can‗t drive 

      Workaround for GECCO-2010:
       Classify the whole track as being straight

      New classifier for the next leg at WCCI-2010




2010 Simulated Car Racing Championship @ GECCO-2010
Jorge Muñoz Fuentes




  Department of Computer Science
  Carlos III University of Madrid

2010 Simulated Car Racing Championship @ GECCO-2010
Jorge Muñoz


      Build a model of the track during the warm-up stage.

      Two neural networks to predict the trajectory using the track
       model. Two neural networks to predict the target speed
       given the model of the track and the current car position

      The four neural networks are trained with backpropagation
       using data retrieved from a human player.

      The controller tries to imitate a human player.

      A scripted policy is used to follow the trajectory (steering
       value), set the speed (accelerate and brake values), set the
       clutch and the gear


2010 Simulated Car Racing Championship @ GECCO-2010
Jorge Muñoz


      Other optimizations performed during
       the warm-up and used in the race:
          The car remember where it goes out of the car or drives
          far form the trajectory and in the next laps goes slower in
          those points
          The car remember where it follows the trajectory perfectly
          and tries to go faster in the next laps.

      Overtaking is made by means of modifying the predicted
       trajectory, the modification is bigger in straighs than in turns

      To avoid overtakin the car also modifies the trajectoy, trying
       to stay in front of the opponent car.




2010 Simulated Car Racing Championship @ GECCO-2010
Joseph Alton




                 Joseph Alton


2010 Simulated Car Racing Championship @ GECCO-2010
Joseph Alton


      Bot made specially for the oval leg, and is only designed to
       be competent on this leg.
      Plan to change and improve the bot as I learn and to meet
       the demands of the other tracks

      Warm-up Learning
         Car goes slowly around the track recording in each
         segment (e.g. 0 metre, 1 metre, 2 metre) whether it is
         turning or not
         This information can then be loaded up in qualifying or
         finals to help find a more appropriate target speed.
         Straight = fast, Turn = slower
         During the race, the driver looks ahead to the next or the
         next two segments depending on the speed



2010 Simulated Car Racing Championship @ GECCO-2010
COBOSTAR




      Thies Lönneker and Martin V. Butz
      University of Würzburg

      http://www.coboslab.psychologie.uni-wuerzburg.de
2010 Simulated Car Racing Championship @ GECCO-2010
CIG-2008 Champ




                                                          Luigi Cardamone
                                                      Politecnico di Milano



2010 Simulated Car Racing Championship @ GECCO-2010
Qualifying
Scoring process: Warm-up Qualifying


      Scoring process involves three motor high-speedway tracks:
          Kiwi
          Mango
          Triangle

      Kiwi is a (renamed and slightly modified) track of TORCS

      Mango and Triangle has been developed by the organizers

      Each controller raced for 100000 game ticks in the warm-up
       stage and then its performance is computed in the qualifying
       stage as the distance covered within 10000 game ticks




2010 Simulated Car Racing Championship @ GECCO-2010
Qualifying: Kiwi




            MR.Racer


              Joseph



          Jorge Muñoz


          COBOSTAR



           Cardamone


              Autopia


                        0   2000   4000   6000   8000   10000   12000   14000   16000   18000




2010 Simulated Car Racing Championship @ GECCO-2010
Qualifying: Mango




            MR.Racer


               Joseph


          Jorge Muñoz


          COBOSTAR


           Cardamone


              Autopia


                        0   2000     4000    6000     8000   10000   12000   14000   16000



2010 Simulated Car Racing Championship @ GECCO-2010
Qualifying: Triangle




            MR.Racer


              Joseph


          Jorge Muñoz


          COBOSTAR


           Cardamone


              Autopia


                        0   2000    4000    6000      8000   10000   12000   14000   16000



2010 Simulated Car Racing Championship @ GECCO-2010
Qualifyng summary




                Competitor           Kiwi       Mango      Triangle   Total
                COBOSTAR              10              10      8        28
                Autopia                8              8      10        26
                Mr. Racer              6              6       6        18
                Jorge Muñoz            4              5       5        14
                Cardamone              5              4       4        13
                Joseph                 3              3       3        9




2010 Simulated Car Racing Championship @ GECCO-2010
What about qualifying?
              COBOSTAR is the fastest driver
(despite not optimized neither for high-speedways nor for
                     noisy sensors)

         Experiments without noise shows that

         Noise is not critical in high-speedways
                    for driving alone

 Some controllers (MR. Racer and Muñoz) perform even
                   better with noise
The GECCO-2010 Leg
35




                                       The Three GPs

                         For each track we run 5 races
                          with random starting grids

           Each race is scored using the F1 point system
              (10 to first, 8 to second, 6 to third, …)

          Two points to the controller with lesser damage

                Two points for the fastest lap of the race



2010 Simulated Car Racing Championship @ GECCO-2010
Final Results



        Competitor                 Kiwi         Mango       Triangle   Total
        Autopia                     12                12      10        34
        Jorge Muñoz                 5.5               8        9       22.5
        Mr. Racer                    9                4        3        16
        Cardamone                    6                4.5     5.5       16
        Joseph                       4                5       6.5      15.5
        COBOSTAR                     4                4.5     5.5       14




2010 Simulated Car Racing Championship @ GECCO-2010
2010 Simulated Car Racing Championship @ GECCO-2010
What about the race?
    Results do not confirm the qualifying performance

      COBOSTAR is the worst performing controller!

The results suggest that noise is much more relevant for the
        opponent perception that for driving alone

  Races without noise show that AUTOPIA and COBOSTAR
   being the top controllers (as expected from qualifyng)

Cardamone’s controller seems slighlty more reliable (as it is
       completely based on a Neural Network)

Weitere ähnliche Inhalte

Ähnlich wie 2010 Simulated Car Racing Championship @ GECCO-2010

2010 Simulated Car Racing Championship @ CIG-2010
2010 Simulated Car Racing Championship @ CIG-20102010 Simulated Car Racing Championship @ CIG-2010
2010 Simulated Car Racing Championship @ CIG-2010
Daniele Loiacono
 
Kenyon Kluge Inboard -- Driving Production Innovation Through Race Technology...
Kenyon Kluge Inboard -- Driving Production Innovation Through Race Technology...Kenyon Kluge Inboard -- Driving Production Innovation Through Race Technology...
Kenyon Kluge Inboard -- Driving Production Innovation Through Race Technology...
Kenyon Kluge
 
Summary of the Simulated Car Racing Optimization Contest@GECCO-09
Summary of the Simulated Car Racing Optimization Contest@GECCO-09Summary of the Simulated Car Racing Optimization Contest@GECCO-09
Summary of the Simulated Car Racing Optimization Contest@GECCO-09
Daniele Loiacono
 
Tobias Overdiek - Resume Presentation 2016.v3
Tobias Overdiek - Resume Presentation 2016.v3Tobias Overdiek - Resume Presentation 2016.v3
Tobias Overdiek - Resume Presentation 2016.v3
Tobias Overdiek
 
Romain AVY CV 2016
Romain AVY CV 2016Romain AVY CV 2016
Romain AVY CV 2016
Romain Avy
 

Ähnlich wie 2010 Simulated Car Racing Championship @ GECCO-2010 (20)

2010 Simulated Car Racing Championship @ CIG-2010
2010 Simulated Car Racing Championship @ CIG-20102010 Simulated Car Racing Championship @ CIG-2010
2010 Simulated Car Racing Championship @ CIG-2010
 
2011 Simulated Car Racing Championship @ GECCO-2011
2011 Simulated Car Racing Championship @ GECCO-20112011 Simulated Car Racing Championship @ GECCO-2011
2011 Simulated Car Racing Championship @ GECCO-2011
 
2009 Simulate Car Racing Championship
2009 Simulate Car Racing Championship2009 Simulate Car Racing Championship
2009 Simulate Car Racing Championship
 
2012 Simulated Car Racing Championship @ GECCO-2012
2012 Simulated Car Racing Championship @ GECCO-20122012 Simulated Car Racing Championship @ GECCO-2012
2012 Simulated Car Racing Championship @ GECCO-2012
 
AUM-Proposal_omar
AUM-Proposal_omarAUM-Proposal_omar
AUM-Proposal_omar
 
The 2009 Simulated Car Racing Championship
The 2009 Simulated Car Racing ChampionshipThe 2009 Simulated Car Racing Championship
The 2009 Simulated Car Racing Championship
 
Regulatory Outlook for real-world emissions - ICCT
Regulatory Outlook for real-world emissions - ICCTRegulatory Outlook for real-world emissions - ICCT
Regulatory Outlook for real-world emissions - ICCT
 
Kenyon Kluge Inboard -- Driving Production Innovation Through Race Technology...
Kenyon Kluge Inboard -- Driving Production Innovation Through Race Technology...Kenyon Kluge Inboard -- Driving Production Innovation Through Race Technology...
Kenyon Kluge Inboard -- Driving Production Innovation Through Race Technology...
 
Equipment Remote Fleet Monitoring System
Equipment Remote Fleet Monitoring SystemEquipment Remote Fleet Monitoring System
Equipment Remote Fleet Monitoring System
 
Summary of the Simulated Car Racing Optimization Contest@GECCO-09
Summary of the Simulated Car Racing Optimization Contest@GECCO-09Summary of the Simulated Car Racing Optimization Contest@GECCO-09
Summary of the Simulated Car Racing Optimization Contest@GECCO-09
 
2014 Mitsubishi Outlander Sport Brochure
2014 Mitsubishi Outlander Sport Brochure2014 Mitsubishi Outlander Sport Brochure
2014 Mitsubishi Outlander Sport Brochure
 
Gt3 open wc denglish
Gt3 open wc denglishGt3 open wc denglish
Gt3 open wc denglish
 
Benchmark Magazine article
Benchmark Magazine articleBenchmark Magazine article
Benchmark Magazine article
 
Improving_programming_skills_of_Mechanical_Enginee.pdf
Improving_programming_skills_of_Mechanical_Enginee.pdfImproving_programming_skills_of_Mechanical_Enginee.pdf
Improving_programming_skills_of_Mechanical_Enginee.pdf
 
Tobias Overdiek - Resume Presentation 2016.v3
Tobias Overdiek - Resume Presentation 2016.v3Tobias Overdiek - Resume Presentation 2016.v3
Tobias Overdiek - Resume Presentation 2016.v3
 
Tucson-brochure-En.pdf
Tucson-brochure-En.pdfTucson-brochure-En.pdf
Tucson-brochure-En.pdf
 
Team eta presentation 2020
Team eta presentation 2020Team eta presentation 2020
Team eta presentation 2020
 
Presentation @ Miniscuola WOA 2015
Presentation @ Miniscuola WOA 2015Presentation @ Miniscuola WOA 2015
Presentation @ Miniscuola WOA 2015
 
Xpeng Motors' P7's self-driving roadmap and system design
Xpeng Motors' P7's self-driving roadmap and system design Xpeng Motors' P7's self-driving roadmap and system design
Xpeng Motors' P7's self-driving roadmap and system design
 
Romain AVY CV 2016
Romain AVY CV 2016Romain AVY CV 2016
Romain AVY CV 2016
 

Mehr von Daniele Loiacono

CIG-2008 Car Racing Competition
CIG-2008 Car Racing CompetitionCIG-2008 Car Racing Competition
CIG-2008 Car Racing Competition
Daniele Loiacono
 
Evolving Trading Rules with XCSF
Evolving Trading Rules with XCSFEvolving Trading Rules with XCSF
Evolving Trading Rules with XCSF
Daniele Loiacono
 
Ant Wars: XCS Strikes Back
Ant Wars: XCS Strikes BackAnt Wars: XCS Strikes Back
Ant Wars: XCS Strikes Back
Daniele Loiacono
 
Evolving Neural Networks Through Augmenting Topologies
Evolving Neural Networks Through Augmenting TopologiesEvolving Neural Networks Through Augmenting Topologies
Evolving Neural Networks Through Augmenting Topologies
Daniele Loiacono
 

Mehr von Daniele Loiacono (20)

GPUs for GEC Competition @ GECCO-2013
GPUs for GEC Competition @ GECCO-2013GPUs for GEC Competition @ GECCO-2013
GPUs for GEC Competition @ GECCO-2013
 
EvoRobocode Competition @ GECCO-2013
EvoRobocode Competition @ GECCO-2013EvoRobocode Competition @ GECCO-2013
EvoRobocode Competition @ GECCO-2013
 
Computational Intelligence in Games Tutorial @GECCO2012
Computational Intelligence in Games Tutorial @GECCO2012Computational Intelligence in Games Tutorial @GECCO2012
Computational Intelligence in Games Tutorial @GECCO2012
 
XCSF with Local Deletion: Preventing Detrimental Forgetting
XCSF with Local Deletion: Preventing Detrimental ForgettingXCSF with Local Deletion: Preventing Detrimental Forgetting
XCSF with Local Deletion: Preventing Detrimental Forgetting
 
Testing learning classifier systems
Testing learning classifier systemsTesting learning classifier systems
Testing learning classifier systems
 
Random Artificial Incorporation of Noise in a Learning Classifier System Envi...
Random Artificial Incorporation of Noise in a Learning Classifier System Envi...Random Artificial Incorporation of Noise in a Learning Classifier System Envi...
Random Artificial Incorporation of Noise in a Learning Classifier System Envi...
 
One Step Fits All
One Step Fits AllOne Step Fits All
One Step Fits All
 
Introducing LCS to Digital Design Verification
Introducing LCS to Digital Design VerificationIntroducing LCS to Digital Design Verification
Introducing LCS to Digital Design Verification
 
A temporal classifier system using spiking neural networks
A temporal classifier system using spiking neural networksA temporal classifier system using spiking neural networks
A temporal classifier system using spiking neural networks
 
Confusion Matrices for Improving Performance of Feature Pattern Classifier Sy...
Confusion Matrices for Improving Performance of Feature Pattern Classifier Sy...Confusion Matrices for Improving Performance of Feature Pattern Classifier Sy...
Confusion Matrices for Improving Performance of Feature Pattern Classifier Sy...
 
Automatically Defined Functions for Learning Classifier Systems
Automatically Defined Functions for Learning Classifier SystemsAutomatically Defined Functions for Learning Classifier Systems
Automatically Defined Functions for Learning Classifier Systems
 
Voting Based Learning Classifier System for Multi-Label Classification
Voting Based Learning Classifier System for Multi-Label ClassificationVoting Based Learning Classifier System for Multi-Label Classification
Voting Based Learning Classifier System for Multi-Label Classification
 
CIG-2008 Car Racing Competition
CIG-2008 Car Racing CompetitionCIG-2008 Car Racing Competition
CIG-2008 Car Racing Competition
 
Car Racing Competition at WCCI2008 - Summary
Car Racing Competition at WCCI2008 - SummaryCar Racing Competition at WCCI2008 - Summary
Car Racing Competition at WCCI2008 - Summary
 
Car Racing Competition at WCCI2008 - Diego Perez
Car Racing Competition at WCCI2008 - Diego PerezCar Racing Competition at WCCI2008 - Diego Perez
Car Racing Competition at WCCI2008 - Diego Perez
 
Cat Racing Competition at WCCI 2008 - Leonard Kinaird Heether
Cat Racing Competition at WCCI 2008 - Leonard Kinaird HeetherCat Racing Competition at WCCI 2008 - Leonard Kinaird Heether
Cat Racing Competition at WCCI 2008 - Leonard Kinaird Heether
 
Car Racing Competition at WCCI2008
Car Racing Competition at WCCI2008Car Racing Competition at WCCI2008
Car Racing Competition at WCCI2008
 
Evolving Trading Rules with XCSF
Evolving Trading Rules with XCSFEvolving Trading Rules with XCSF
Evolving Trading Rules with XCSF
 
Ant Wars: XCS Strikes Back
Ant Wars: XCS Strikes BackAnt Wars: XCS Strikes Back
Ant Wars: XCS Strikes Back
 
Evolving Neural Networks Through Augmenting Topologies
Evolving Neural Networks Through Augmenting TopologiesEvolving Neural Networks Through Augmenting Topologies
Evolving Neural Networks Through Augmenting Topologies
 

2010 Simulated Car Racing Championship @ GECCO-2010

  • 1. The 2010 Simulated Car Racing Championship @ GECCO-2010 Daniele Loiacono, Luigi Cardamone, Martin V. Butz, and Pier Luca Lanzi
  • 2. 2 2010 Simulated Car Racing Championship 9 races during 3 conferences ACM GECCO-2010, Portland, OR (USA), July 7- IEEE WCCI-2010, Barcelona (Spain), July 18-23 IEEE CIG-2010, Copenhagen (Denmark), August 18–21 Develop a driver for TORCS (hand-coded, learned, evolved, …) Drivers will be awarded based on their score in each conference competition At the end, the team with highest overall score wins the championship 2010 Simulated Car Racing Championship @ GECCO-2010
  • 3. 3 What is the structure of a race? Three stages: warm up, qualifiers, actual race During warm-up, each driver can explore the track and learn something useful During qualifiers, each driver races alone against the clock (the best 8 drivers move to the race) During the race all the drivers race together 2010 Simulated Car Racing Championship @ GECCO-2010
  • 4. Motivations  Proposing a relevant game-based competition more representative of commercial games AI more similar to a real-world problem  Proposing a funny and exciting competition you can see and play with the entries of this competition human players can interact with AI a lot of programmed AI available for comparison  Proposing a challenging competition not designed with Machine Learning in mind computationally expensive real-time dealing with a lot of practical issues 2010 Simulated Car Racing Championship @ GECCO-2010
  • 5. What‘s new? If everything seems under control, you're not going fast enough — Mario Andretti  Warm-up stage Before qualifying stage, competitors have 100000 game ticks to race on the track Allows track learning and optimization of parameters  Noisy sensors Track sensors and opponent sensors are affected by a Gaussian noise (standard deviation equal to 10% of the readings)  Extended sensor model Focus sensors Z position and speed Direction of track sensors fully customizable Added clutch control and focus command 2010 Simulated Car Racing Championship @ GECCO-2010
  • 7. The Open Racing Car Simulator  TORCS is a state of the art open source simulator written in C++  Main features Sophisticated dynamics Provided with several cars, tracks, and controllers Active community of users and developers Easy to develop your own controller  OS Support Linux: binaries and building from sources Windows: binaries and ―limited‖ bulding from sources support OSX: legacy binaries and no building from sources support  2010 Simulated Car Racing Championship @ GECCO-2010
  • 8. The Open Racing Car Simulator & the Competition Software TORCS TORCS PATCH BOT BOT BOT SBOT SBOT SBOT  The competition server UDP UDP UDP  Separates the bots from TORCS  Build a well-defined sensor model  Works in real-time BOT BOT BOT 2010 Simulated Car Racing Championship @ GECCO-2010
  • 9. Sensors and actuators  Rangefinders for edges on the track and opponents  Speed, RPM, fuel, damage, angle with track, distance race, position on track, etc.  Four effectors: steering wheel [-1,+1], gas pedal [0, +1], Brake pedal [0,+1], Gearbox {-1,0,1,2,3,4,5,6} 2010 Simulated Car Racing Championship @ GECCO-2010
  • 11. The competitors  Four entries submitted to this first leg AUTOPIA, Madrid and Granada J. Muñoz, Carlos III University of Madrid S.Pohl, J. Quadflieg and T. Delbrügger, TU Dortmund Joseph Alton, University of Birmingham  Two more entries from the 2009 championship COBOSTAR (T. Lönneker & M.V. Butz, University of Würzburg) POLIMI (Cardamone, Politecnico di Milano) 2010 Simulated Car Racing Championship @ GECCO-2010
  • 12. AUTOPIA Industrial Computer Science Department. Centro de Automática y Robótica Consejo Superior de Investigaciones Científicas Madrid, Spain Contact:E. Onieva (enrique.onieva@car.upm-csic.es) 2010 Simulated Car Racing Championship @ GECCO-2010
  • 13. Architecture Schema  Three basic modules for gear, steering and speed control  Steady stage genetic algorithm to compute the best weights to combine parameters for steering and target speed control  Opponents module Acts on steering and brake signal to overtake opponents and avoid collisions  Learning Module in Warm-up Stage Factors over the target speed in certain track segments 2010 Simulated Car Racing Championship @ GECCO-2010
  • 14. Opponents Management (Steering)  If O(-20,-10,10,20) detects an opponent with lateral displacement less than 20 meters, move the steering trackwidth/100 through the direction with more free distance.  If O(-40,-30,-20,-10) detects an opponent with lateral displacement less than 20 meters, move the steering trackwidth/50 through the right.  If O(-70,-60,-50)<15 meters, move the steering trackwidth/50 through the right.  If O(-140,-130,-120,-110,-100,-90,-80)<15meters , move the steering trackwidth/50 through the right.  Mirroring actions for positive sensors  O(X)  Opponent Measure of Sensor Oriented at X  Steering Modification limited to 0.35 2010 Simulated Car Racing Championship @ GECCO-2010
  • 15. Learning Module (Warm-Up)  Running normally in warm-up stage.  Maintain a vector with as many real values as the track length in meters.  Vector initialized to 1.0  If the vehicle goes out of the track or suffers damage then multiply vector positions from 100 meters before the current position by 0.9. 2010 Simulated Car Racing Championship @ GECCO-2010
  • 16. Mr Racer Susanna Pohl, Jan Quadflieg and Tim Delbrügger TU Dortmund 2010 Simulated Car Racing Championship @ GECCO-2010
  • 17. Mr. Racer 2009-2010  Mr Racer 2009 Good classifier which identifies six situations Acceleration/brake learned offline using an EA Model of the track learned online Simple heuristic to use the model: override the learned behaviour on straights and in full speed corners to drive flat out  Mr Racer 2010 Save the model after warmup, use it during qualifying and the race Use the model to derive a plan consisting of target speeds and a racing line Optimize the parameter set of the planning module, left to be done for the next round 2010 Simulated Car Racing Championship @ GECCO-2010
  • 18. Mr. Racer - Classification Angle based measure mapped to six situations (straight, fast corner, slow corner, etc) 2010 Simulated Car Racing Championship @ GECCO-2010
  • 19. Mr. Racer 2010 – The catch  Noise completely breaks our classifier  Without a descent classifier we can‗t learn the track  Without a trackmodel we can‗t drive   Workaround for GECCO-2010: Classify the whole track as being straight  New classifier for the next leg at WCCI-2010 2010 Simulated Car Racing Championship @ GECCO-2010
  • 20. Jorge Muñoz Fuentes Department of Computer Science Carlos III University of Madrid 2010 Simulated Car Racing Championship @ GECCO-2010
  • 21. Jorge Muñoz  Build a model of the track during the warm-up stage.  Two neural networks to predict the trajectory using the track model. Two neural networks to predict the target speed given the model of the track and the current car position  The four neural networks are trained with backpropagation using data retrieved from a human player.  The controller tries to imitate a human player.  A scripted policy is used to follow the trajectory (steering value), set the speed (accelerate and brake values), set the clutch and the gear 2010 Simulated Car Racing Championship @ GECCO-2010
  • 22. Jorge Muñoz  Other optimizations performed during the warm-up and used in the race: The car remember where it goes out of the car or drives far form the trajectory and in the next laps goes slower in those points The car remember where it follows the trajectory perfectly and tries to go faster in the next laps.  Overtaking is made by means of modifying the predicted trajectory, the modification is bigger in straighs than in turns  To avoid overtakin the car also modifies the trajectoy, trying to stay in front of the opponent car. 2010 Simulated Car Racing Championship @ GECCO-2010
  • 23. Joseph Alton Joseph Alton 2010 Simulated Car Racing Championship @ GECCO-2010
  • 24. Joseph Alton  Bot made specially for the oval leg, and is only designed to be competent on this leg.  Plan to change and improve the bot as I learn and to meet the demands of the other tracks  Warm-up Learning Car goes slowly around the track recording in each segment (e.g. 0 metre, 1 metre, 2 metre) whether it is turning or not This information can then be loaded up in qualifying or finals to help find a more appropriate target speed. Straight = fast, Turn = slower During the race, the driver looks ahead to the next or the next two segments depending on the speed 2010 Simulated Car Racing Championship @ GECCO-2010
  • 25. COBOSTAR Thies Lönneker and Martin V. Butz University of Würzburg http://www.coboslab.psychologie.uni-wuerzburg.de 2010 Simulated Car Racing Championship @ GECCO-2010
  • 26. CIG-2008 Champ Luigi Cardamone Politecnico di Milano 2010 Simulated Car Racing Championship @ GECCO-2010
  • 28. Scoring process: Warm-up Qualifying  Scoring process involves three motor high-speedway tracks: Kiwi Mango Triangle  Kiwi is a (renamed and slightly modified) track of TORCS  Mango and Triangle has been developed by the organizers  Each controller raced for 100000 game ticks in the warm-up stage and then its performance is computed in the qualifying stage as the distance covered within 10000 game ticks 2010 Simulated Car Racing Championship @ GECCO-2010
  • 29. Qualifying: Kiwi MR.Racer Joseph Jorge Muñoz COBOSTAR Cardamone Autopia 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 2010 Simulated Car Racing Championship @ GECCO-2010
  • 30. Qualifying: Mango MR.Racer Joseph Jorge Muñoz COBOSTAR Cardamone Autopia 0 2000 4000 6000 8000 10000 12000 14000 16000 2010 Simulated Car Racing Championship @ GECCO-2010
  • 31. Qualifying: Triangle MR.Racer Joseph Jorge Muñoz COBOSTAR Cardamone Autopia 0 2000 4000 6000 8000 10000 12000 14000 16000 2010 Simulated Car Racing Championship @ GECCO-2010
  • 32. Qualifyng summary Competitor Kiwi Mango Triangle Total COBOSTAR 10 10 8 28 Autopia 8 8 10 26 Mr. Racer 6 6 6 18 Jorge Muñoz 4 5 5 14 Cardamone 5 4 4 13 Joseph 3 3 3 9 2010 Simulated Car Racing Championship @ GECCO-2010
  • 33. What about qualifying? COBOSTAR is the fastest driver (despite not optimized neither for high-speedways nor for noisy sensors) Experiments without noise shows that Noise is not critical in high-speedways for driving alone Some controllers (MR. Racer and Muñoz) perform even better with noise
  • 35. 35 The Three GPs For each track we run 5 races with random starting grids Each race is scored using the F1 point system (10 to first, 8 to second, 6 to third, …) Two points to the controller with lesser damage Two points for the fastest lap of the race 2010 Simulated Car Racing Championship @ GECCO-2010
  • 36. Final Results Competitor Kiwi Mango Triangle Total Autopia 12 12 10 34 Jorge Muñoz 5.5 8 9 22.5 Mr. Racer 9 4 3 16 Cardamone 6 4.5 5.5 16 Joseph 4 5 6.5 15.5 COBOSTAR 4 4.5 5.5 14 2010 Simulated Car Racing Championship @ GECCO-2010
  • 37. 2010 Simulated Car Racing Championship @ GECCO-2010
  • 38. What about the race? Results do not confirm the qualifying performance COBOSTAR is the worst performing controller! The results suggest that noise is much more relevant for the opponent perception that for driving alone Races without noise show that AUTOPIA and COBOSTAR being the top controllers (as expected from qualifyng) Cardamone’s controller seems slighlty more reliable (as it is completely based on a Neural Network)