SlideShare a Scribd company logo
Beating uncertainty
in racing bot evolution through
enhanced exploration and pole
position selection
Mohammed Salem (University of Mascara, Algeria)
Antonio M. Mora (University of Granada, Spain)
J.J. Merelo (University of Granada, Spain)
IEEE Conference on Games 2019
20-23 August
London, UK
Index
Beating uncertainty
in racing bot
evolution
through
enhanced exploration
and
pole position
selection
The Open Racing Car Simulator.
Used in Competitions and Testbed for AI
Controllers (Driving bots)
Inputs/Outputs:
– Sensors: Car/Track distances
– Actuators: Steer, Break,
Accelerator, Gear
http://torcs.sourceforge.net/
TORCS
Racing bot
TORCS
Racing bot
Controller Architecture
Racing bot
Target Speed Fuzzy Subcontroller
Racing bot
Mamdani Fuzzy System with
3 Trapezoidal Membership
Functions (Low, Medium, High).
Target Speed Fuzzy Subcontroller
Racing bot
IF Front = High THEN TargetSpeed is 280
IF Front = Medium THEN TargetSpeed is 240
IF Front = Low AND M5 = High THEN TargetSpeed is 220
IF Front = Low AND M5 = Medium THEN TargetSpeed is 180
IF Front = Low AND M5 = Low AND M10 = High THEN TargetSpeed is 120
IF Front = Low AND M5 = Low AND M10 = Medium THEN TargetSpeed is 60
IF Front = Low AND M5 = Low AND M10 = Low THEN TargetSpeed is 30
IF Front = MAXDISTSPEED OR M5 = MAXDISTSPEED OR M10 =
MAXDISTSPEED THEN TargetSpeed = 300
Uses a Set of Fuzzy Rules:
Steer Control Fuzzy Subcontroller
Racing bot
Mamdani Fuzzy System with 3
Trapezoidal Membership
Functions.
Same inputs (FRONT, M5 and M10).
Set of Fuzzy Rules:
IF Front = High THEN steer = 0
IF Front = Medium AND M10 = High THEN steer = 0.25
IF Front = Medium AND M10 = Medium AND M5 = Medium THEN steer = 0.25
IF Front = Medium AND M10 = Medium AND M5 = Low THEN steer = 0.5
IF Front = Low AND M10 = High THEN steer = 0.5
IF Front = Low AND M10 = Medium AND M5 = Medium THEN steer = 0.75
IF Front = Low AND M10 = Medium AND M5 = Low THEN steer = 0.75
Beating uncertainty
in racing bot
evolution
through
enhanced exploration
and
pole position
selection
Index
Genetic Algorithm
evolution
Wikimedia
Genetic Algorithm
evolution
Population: set of controllers defined each one by
a vector which determines the shape of the
membership functions.
Evaluation: every
candidate controller must
be evaluated in the
environment (TORCS races).
A value must be assigned
to its performance =>
Fitness Function.
Genetic Algorithm
evolution
Fitness Function:
- Maximize average speed on the whole circuit while
minimizing damage.
- Focuses on objectives of a (human) driver during a
race, not just on winning or losing => human-like.
Genetic Algorithm
evolution
Evaluation process:
AVG(Speed)
Damage
Beating uncertainty
in racing bot
evolution
through
enhanced exploration
and
pole position
selection
Index
Challenges
uncertainty
Reliable and effective
evaluation process:
 Good individuals <=>
Competitive controllers.
 Selection of the best.
Variable results for the
same controller:
 Weather conditions.
 Track status.
 Non-deterministic behavior of rivals.
Solutions
uncertainty
Fitness definition and
evaluation:
 Considering the main factors
(just speed and damage).
 Several races.
 Several laps.
 No rivals.
 Track combining difficult and
easy parts.
Alpine 2 Track
Beating uncertainty
in racing bot
evolution
through
enhanced exploration
and
pole position
selection
Index
Enhanced
Exploration
Blend Crossover Operator (BLX-a):
 Generates different descendants with random genes.
 The differences (depend on a value).
 If a > 0.5 => higher exploration
 If a < 0.5 => higher exploitation
Two approaches:
 a constant
 a variable
Beating uncertainty
in racing bot
evolution
through
enhanced exploration
and
pole position
selection
Index
Pole position
selection
Or race-based selection:
 Individuals arranged in groups
of 10.
 They compete in several races of
several laps (same car).
 They obtain a F1-based score
depending on their final
position in every race.
 At the end, those 5 with the
best global score are selected
as parents.
Advantages:
 Less sensitive to uncertainty (fitness-less selection)
 Selected individuals will be almost surely the best.
Beating uncertainty
in racing bot
evolution
through
enhanced exploration
and
pole position
selection
Index
Experimental Setup
beating
GA parameters:
 Generations = 50
 Population size = 60
 Crossover rate = 0.85
 Mutation rate = 0.09
Genetic operators:
 Random initialization
 Race-based selection [every 5 generations]
Tournament selection [the rest]
 Non-uniform mutation
 BLX-a crossover
- CONSTANT => a = 0.5
- VARIABLE => a = 1-(generation/TOTAL_GENS)
Evaluation:
 Alpine 2 Track, 20 laps, Solo race, car1-tbr1 (moderate performance)
Selection of the best
beating
At the end of GA run:
 10 best individuals
(according to fitness).
 5 races.
 5 laps.
 Alpine 2 Track.
 Random starting grid.
 F1-based score: 1st - 25 points, 2nd - 18, 3rd - 15,
4 - 12, 5 - 10, 6 - 8, 7 - 6, 8 - 4, 9 - 2, 10 – 1.
At the end of 10 runs:
 The 10 best individuals (selected before).
 5 races/5 laps in Alpine 2.
 Random starting grid.
Results
beating
Facts:
 New approaches improve previous controllers.
 GFC-VA clearly outperforms all the rest.
 Excellent performance even in an unknown track (E-Track 5).
Controllers:
 GFC -> our previous best controller (AVS Fitness).
 GFC-RS -> our previous controller with race-based selection.
 GFC-FA -> new controller with race-based selection and BLX-a
with constant value.
 GFC-VA -> new controller with race-based selection and BLX-a
with variable value.
Results
beating
Facts:
 New approaches improve previous controllers.
 GFC-VA clearly outperforms all the rest.
 Excellent performance even in an unknown track (E-Track 5).
Controllers:
 GFC -> our previous best controller (AVS Fitness).
 GFC-RS -> our previous controller with race-based selection.
 GFC-FA -> new controller with race-based selection and BLX-a
with constant value.
 GFC-VA -> new controller with race-based selection and BLX-a
with variable value.
PSRI Controller
beating
New Challenger: PSRI
 Controller created by Pérez-
Liébana, Recio, Sáez and
Isasi.
 4th in Simulated Car Racing
Competition 2009.
 Finite State Machine using
fuzzy rules for transitions.
 Optimized by means of NSGA-
II.
 J48 classifier for prediction
of parts of the track.
Results
beating
Facts:
 Both GFC-FA and GFC-VA improve all PSRI results.
 PSRI is able to win one of the races on each track.
 GFC-VA is the best of the three (wins 4 and 3 races).
5 races / 20 laps per track. Average results.
We have
Future Work:
 Implement completely Race-based selection (in all the
generations).
 Optimize the output of the fuzzy rules.
 Optimize the set of fuzzy rules (Genetic Programming).
Beaten uncertainty
in racing bot evolution
through
enhanced exploration and
pole position selection
conclusions
amorag@ugr.es
@amoragar

More Related Content

Similar to Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection (CoG 2019)

Present car racing_setup
Present car racing_setupPresent car racing_setup
Present car racing_setupkeldon_spain
 
Application of Artificial Intelligence for Automotive Applications
Application of Artificial Intelligence for Automotive ApplicationsApplication of Artificial Intelligence for Automotive Applications
Application of Artificial Intelligence for Automotive Applications
KonfHubTechConferenc
 
SideSlip Angle Estimator (SSE)
SideSlip Angle Estimator (SSE)SideSlip Angle Estimator (SSE)
SideSlip Angle Estimator (SSE)
Enrico Busto
 
Safe Driving Advisor and Evaluator.pptx
Safe Driving Advisor and Evaluator.pptxSafe Driving Advisor and Evaluator.pptx
Safe Driving Advisor and Evaluator.pptx
Hazem Fahmy
 
China NCAP Updates
China NCAP UpdatesChina NCAP Updates
China NCAP Updates
Global NCAP
 
Encoding Robotic Sensor States for Q-Learning using the
Encoding Robotic Sensor States for Q-Learning using the Encoding Robotic Sensor States for Q-Learning using the
Encoding Robotic Sensor States for Q-Learning using the butest
 
Update on Vehicle Safety Testing Plans at IIHS, Adrian Lund, President IIHS &...
Update on Vehicle Safety Testing Plans at IIHS, Adrian Lund, President IIHS &...Update on Vehicle Safety Testing Plans at IIHS, Adrian Lund, President IIHS &...
Update on Vehicle Safety Testing Plans at IIHS, Adrian Lund, President IIHS &...
Global NCAP
 
Backjin lee autonomous vehicle slideshare
Backjin lee autonomous vehicle slideshareBackjin lee autonomous vehicle slideshare
Backjin lee autonomous vehicle slideshare
Backjin Lee
 
Matthew Avery speaks at Global NCAP Annual Forum 2014 Melbourne
Matthew Avery speaks at Global NCAP Annual Forum 2014 MelbourneMatthew Avery speaks at Global NCAP Annual Forum 2014 Melbourne
Matthew Avery speaks at Global NCAP Annual Forum 2014 Melbourne
Global NCAP
 
Deep_Learning__INAF_baroncelli.pdf
Deep_Learning__INAF_baroncelli.pdfDeep_Learning__INAF_baroncelli.pdf
Deep_Learning__INAF_baroncelli.pdf
asdfasdf214078
 
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
Daniele Loiacono
 
Four Wheel Active Steering / Without Videos
Four Wheel Active Steering / Without VideosFour Wheel Active Steering / Without Videos
Four Wheel Active Steering / Without Videos
Goodarz Mehr
 
Performance Traction Control (PTC)
Performance Traction Control (PTC)Performance Traction Control (PTC)
Performance Traction Control (PTC)
Enrico Busto
 
Algoritmos Genéticos aplicados em Machine Learning
Algoritmos Genéticos aplicados em Machine LearningAlgoritmos Genéticos aplicados em Machine Learning
Algoritmos Genéticos aplicados em Machine Learningbutest
 
Study on Genetic Algorithm Approaches to Improve an Autonomous Agent for a Fi...
Study on Genetic Algorithm Approaches to Improve an Autonomous Agent for a Fi...Study on Genetic Algorithm Approaches to Improve an Autonomous Agent for a Fi...
Study on Genetic Algorithm Approaches to Improve an Autonomous Agent for a Fi...
Antonio Mora
 
Captronics Systems Automotive Testing
Captronics Systems Automotive TestingCaptronics Systems Automotive Testing
Captronics Systems Automotive Testing
Captronic Systems
 
Baja Falcons Tuning and Instrumentation Team IGEN430 Poster - VR (1)
Baja Falcons Tuning and Instrumentation Team IGEN430 Poster - VR (1)Baja Falcons Tuning and Instrumentation Team IGEN430 Poster - VR (1)
Baja Falcons Tuning and Instrumentation Team IGEN430 Poster - VR (1)Vaughn Richards
 
Improving time to-collision estimation by IMM based Kalman filter
Improving time to-collision estimation by IMM based Kalman filterImproving time to-collision estimation by IMM based Kalman filter
Improving time to-collision estimation by IMM based Kalman filter
Yixin Chen
 

Similar to Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection (CoG 2019) (20)

Present car racing_setup
Present car racing_setupPresent car racing_setup
Present car racing_setup
 
Application of Artificial Intelligence for Automotive Applications
Application of Artificial Intelligence for Automotive ApplicationsApplication of Artificial Intelligence for Automotive Applications
Application of Artificial Intelligence for Automotive Applications
 
SideSlip Angle Estimator (SSE)
SideSlip Angle Estimator (SSE)SideSlip Angle Estimator (SSE)
SideSlip Angle Estimator (SSE)
 
Safe Driving Advisor and Evaluator.pptx
Safe Driving Advisor and Evaluator.pptxSafe Driving Advisor and Evaluator.pptx
Safe Driving Advisor and Evaluator.pptx
 
China NCAP Updates
China NCAP UpdatesChina NCAP Updates
China NCAP Updates
 
Encoding Robotic Sensor States for Q-Learning using the
Encoding Robotic Sensor States for Q-Learning using the Encoding Robotic Sensor States for Q-Learning using the
Encoding Robotic Sensor States for Q-Learning using the
 
Update on Vehicle Safety Testing Plans at IIHS, Adrian Lund, President IIHS &...
Update on Vehicle Safety Testing Plans at IIHS, Adrian Lund, President IIHS &...Update on Vehicle Safety Testing Plans at IIHS, Adrian Lund, President IIHS &...
Update on Vehicle Safety Testing Plans at IIHS, Adrian Lund, President IIHS &...
 
Resume -Final
Resume -FinalResume -Final
Resume -Final
 
Backjin lee autonomous vehicle slideshare
Backjin lee autonomous vehicle slideshareBackjin lee autonomous vehicle slideshare
Backjin lee autonomous vehicle slideshare
 
Matthew Avery speaks at Global NCAP Annual Forum 2014 Melbourne
Matthew Avery speaks at Global NCAP Annual Forum 2014 MelbourneMatthew Avery speaks at Global NCAP Annual Forum 2014 Melbourne
Matthew Avery speaks at Global NCAP Annual Forum 2014 Melbourne
 
Deep_Learning__INAF_baroncelli.pdf
Deep_Learning__INAF_baroncelli.pdfDeep_Learning__INAF_baroncelli.pdf
Deep_Learning__INAF_baroncelli.pdf
 
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
 
Four Wheel Active Steering / Without Videos
Four Wheel Active Steering / Without VideosFour Wheel Active Steering / Without Videos
Four Wheel Active Steering / Without Videos
 
Performance Traction Control (PTC)
Performance Traction Control (PTC)Performance Traction Control (PTC)
Performance Traction Control (PTC)
 
Algoritmos Genéticos aplicados em Machine Learning
Algoritmos Genéticos aplicados em Machine LearningAlgoritmos Genéticos aplicados em Machine Learning
Algoritmos Genéticos aplicados em Machine Learning
 
Study on Genetic Algorithm Approaches to Improve an Autonomous Agent for a Fi...
Study on Genetic Algorithm Approaches to Improve an Autonomous Agent for a Fi...Study on Genetic Algorithm Approaches to Improve an Autonomous Agent for a Fi...
Study on Genetic Algorithm Approaches to Improve an Autonomous Agent for a Fi...
 
Captronics Systems Automotive Testing
Captronics Systems Automotive TestingCaptronics Systems Automotive Testing
Captronics Systems Automotive Testing
 
Baja Falcons Tuning and Instrumentation Team IGEN430 Poster - VR (1)
Baja Falcons Tuning and Instrumentation Team IGEN430 Poster - VR (1)Baja Falcons Tuning and Instrumentation Team IGEN430 Poster - VR (1)
Baja Falcons Tuning and Instrumentation Team IGEN430 Poster - VR (1)
 
Session 50Josef Nilsson
Session 50Josef NilssonSession 50Josef Nilsson
Session 50Josef Nilsson
 
Improving time to-collision estimation by IMM based Kalman filter
Improving time to-collision estimation by IMM based Kalman filterImproving time to-collision estimation by IMM based Kalman filter
Improving time to-collision estimation by IMM based Kalman filter
 

More from Antonio Mora

Optimización Adaptativa basada en Colonias de Hormigas para la Composición de...
Optimización Adaptativa basada en Colonias de Hormigas para la Composición de...Optimización Adaptativa basada en Colonias de Hormigas para la Composición de...
Optimización Adaptativa basada en Colonias de Hormigas para la Composición de...
Antonio Mora
 
Adaptive Ant Colony Optimization for Service Function Chaining in a Dynamic 5...
Adaptive Ant Colony Optimization for Service Function Chaining in a Dynamic 5...Adaptive Ant Colony Optimization for Service Function Chaining in a Dynamic 5...
Adaptive Ant Colony Optimization for Service Function Chaining in a Dynamic 5...
Antonio Mora
 
Research in Videogames. (Much) further than just AI
Research in Videogames. (Much) further than just AIResearch in Videogames. (Much) further than just AI
Research in Videogames. (Much) further than just AI
Antonio Mora
 
GRETIVE: Un Bot Evolutivo para HearthStone basado en Perfiles
GRETIVE: Un Bot Evolutivo para HearthStone basado en PerfilesGRETIVE: Un Bot Evolutivo para HearthStone basado en Perfiles
GRETIVE: Un Bot Evolutivo para HearthStone basado en Perfiles
Antonio Mora
 
Checking the difficulty of evolutionary-generated maps in a N-Body inspired m...
Checking the difficulty of evolutionary-generated maps in a N-Body inspired m...Checking the difficulty of evolutionary-generated maps in a N-Body inspired m...
Checking the difficulty of evolutionary-generated maps in a N-Body inspired m...
Antonio Mora
 
Applying Ant Colony Optimization for Service Function Chaining in a 5G Networ...
Applying Ant Colony Optimization for Service Function Chaining in a 5G Networ...Applying Ant Colony Optimization for Service Function Chaining in a 5G Networ...
Applying Ant Colony Optimization for Service Function Chaining in a 5G Networ...
Antonio Mora
 
Investigación en videojuegos. (mucho) Mas allá de la IA
Investigación en videojuegos. (mucho) Mas allá de la IAInvestigación en videojuegos. (mucho) Mas allá de la IA
Investigación en videojuegos. (mucho) Mas allá de la IA
Antonio Mora
 
Inteligencia Computacional en Videojuegos (Meetup GranadAI 2019)
Inteligencia Computacional en Videojuegos (Meetup GranadAI 2019)Inteligencia Computacional en Videojuegos (Meetup GranadAI 2019)
Inteligencia Computacional en Videojuegos (Meetup GranadAI 2019)
Antonio Mora
 
Free Form Evolution for Angry Birds Level Generation
Free Form Evolution for Angry Birds Level GenerationFree Form Evolution for Angry Birds Level Generation
Free Form Evolution for Angry Birds Level Generation
Antonio Mora
 
Ciencia y Videojuegos (ULP 2019)
Ciencia y Videojuegos (ULP 2019)Ciencia y Videojuegos (ULP 2019)
Ciencia y Videojuegos (ULP 2019)
Antonio Mora
 
Predicción de Quiebra Financiera de Empresas Mediante Equilibrado de Datos y ...
Predicción de Quiebra Financiera de Empresas Mediante Equilibrado de Datos y ...Predicción de Quiebra Financiera de Empresas Mediante Equilibrado de Datos y ...
Predicción de Quiebra Financiera de Empresas Mediante Equilibrado de Datos y ...
Antonio Mora
 
Sólo puede quedar uno: Evolución de Bots para RTS basada en supervivencia
Sólo puede quedar uno: Evolución de Bots para RTS basada en supervivenciaSólo puede quedar uno: Evolución de Bots para RTS basada en supervivencia
Sólo puede quedar uno: Evolución de Bots para RTS basada en supervivencia
Antonio Mora
 
Living-UGR: Una aventura gráfica geolocalizada para difundir el patrimonio de...
Living-UGR: Una aventura gráfica geolocalizada para difundir el patrimonio de...Living-UGR: Una aventura gráfica geolocalizada para difundir el patrimonio de...
Living-UGR: Una aventura gráfica geolocalizada para difundir el patrimonio de...
Antonio Mora
 
Gamification in Teaching - How to motivate students through games
Gamification in Teaching - How to motivate students through gamesGamification in Teaching - How to motivate students through games
Gamification in Teaching - How to motivate students through games
Antonio Mora
 
Ciencia y videojuegos (versión Extracción de Información) [UCA 05/2015]
Ciencia y videojuegos (versión Extracción de Información) [UCA 05/2015]Ciencia y videojuegos (versión Extracción de Información) [UCA 05/2015]
Ciencia y videojuegos (versión Extracción de Información) [UCA 05/2015]
Antonio Mora
 
Ciencia y Videojuegos (v2)
Ciencia y Videojuegos (v2)Ciencia y Videojuegos (v2)
Ciencia y Videojuegos (v2)
Antonio Mora
 
Evolving Evil: Optimizing Flocking Strategies through Genetic Algorithms for ...
Evolving Evil: Optimizing Flocking Strategies through Genetic Algorithms for ...Evolving Evil: Optimizing Flocking Strategies through Genetic Algorithms for ...
Evolving Evil: Optimizing Flocking Strategies through Genetic Algorithms for ...
Antonio Mora
 
MUSES: A Corporate User-Centric System which Applies Computational Intelligen...
MUSES: A Corporate User-Centric System which Applies Computational Intelligen...MUSES: A Corporate User-Centric System which Applies Computational Intelligen...
MUSES: A Corporate User-Centric System which Applies Computational Intelligen...
Antonio Mora
 
Publicación de imágenes en la web
Publicación de imágenes en la webPublicación de imágenes en la web
Publicación de imágenes en la web
Antonio Mora
 
Modelling Human Expert Behaviour in an Unreal Tournament 2004 Bot
Modelling Human Expert Behaviour in an Unreal Tournament 2004 BotModelling Human Expert Behaviour in an Unreal Tournament 2004 Bot
Modelling Human Expert Behaviour in an Unreal Tournament 2004 Bot
Antonio Mora
 

More from Antonio Mora (20)

Optimización Adaptativa basada en Colonias de Hormigas para la Composición de...
Optimización Adaptativa basada en Colonias de Hormigas para la Composición de...Optimización Adaptativa basada en Colonias de Hormigas para la Composición de...
Optimización Adaptativa basada en Colonias de Hormigas para la Composición de...
 
Adaptive Ant Colony Optimization for Service Function Chaining in a Dynamic 5...
Adaptive Ant Colony Optimization for Service Function Chaining in a Dynamic 5...Adaptive Ant Colony Optimization for Service Function Chaining in a Dynamic 5...
Adaptive Ant Colony Optimization for Service Function Chaining in a Dynamic 5...
 
Research in Videogames. (Much) further than just AI
Research in Videogames. (Much) further than just AIResearch in Videogames. (Much) further than just AI
Research in Videogames. (Much) further than just AI
 
GRETIVE: Un Bot Evolutivo para HearthStone basado en Perfiles
GRETIVE: Un Bot Evolutivo para HearthStone basado en PerfilesGRETIVE: Un Bot Evolutivo para HearthStone basado en Perfiles
GRETIVE: Un Bot Evolutivo para HearthStone basado en Perfiles
 
Checking the difficulty of evolutionary-generated maps in a N-Body inspired m...
Checking the difficulty of evolutionary-generated maps in a N-Body inspired m...Checking the difficulty of evolutionary-generated maps in a N-Body inspired m...
Checking the difficulty of evolutionary-generated maps in a N-Body inspired m...
 
Applying Ant Colony Optimization for Service Function Chaining in a 5G Networ...
Applying Ant Colony Optimization for Service Function Chaining in a 5G Networ...Applying Ant Colony Optimization for Service Function Chaining in a 5G Networ...
Applying Ant Colony Optimization for Service Function Chaining in a 5G Networ...
 
Investigación en videojuegos. (mucho) Mas allá de la IA
Investigación en videojuegos. (mucho) Mas allá de la IAInvestigación en videojuegos. (mucho) Mas allá de la IA
Investigación en videojuegos. (mucho) Mas allá de la IA
 
Inteligencia Computacional en Videojuegos (Meetup GranadAI 2019)
Inteligencia Computacional en Videojuegos (Meetup GranadAI 2019)Inteligencia Computacional en Videojuegos (Meetup GranadAI 2019)
Inteligencia Computacional en Videojuegos (Meetup GranadAI 2019)
 
Free Form Evolution for Angry Birds Level Generation
Free Form Evolution for Angry Birds Level GenerationFree Form Evolution for Angry Birds Level Generation
Free Form Evolution for Angry Birds Level Generation
 
Ciencia y Videojuegos (ULP 2019)
Ciencia y Videojuegos (ULP 2019)Ciencia y Videojuegos (ULP 2019)
Ciencia y Videojuegos (ULP 2019)
 
Predicción de Quiebra Financiera de Empresas Mediante Equilibrado de Datos y ...
Predicción de Quiebra Financiera de Empresas Mediante Equilibrado de Datos y ...Predicción de Quiebra Financiera de Empresas Mediante Equilibrado de Datos y ...
Predicción de Quiebra Financiera de Empresas Mediante Equilibrado de Datos y ...
 
Sólo puede quedar uno: Evolución de Bots para RTS basada en supervivencia
Sólo puede quedar uno: Evolución de Bots para RTS basada en supervivenciaSólo puede quedar uno: Evolución de Bots para RTS basada en supervivencia
Sólo puede quedar uno: Evolución de Bots para RTS basada en supervivencia
 
Living-UGR: Una aventura gráfica geolocalizada para difundir el patrimonio de...
Living-UGR: Una aventura gráfica geolocalizada para difundir el patrimonio de...Living-UGR: Una aventura gráfica geolocalizada para difundir el patrimonio de...
Living-UGR: Una aventura gráfica geolocalizada para difundir el patrimonio de...
 
Gamification in Teaching - How to motivate students through games
Gamification in Teaching - How to motivate students through gamesGamification in Teaching - How to motivate students through games
Gamification in Teaching - How to motivate students through games
 
Ciencia y videojuegos (versión Extracción de Información) [UCA 05/2015]
Ciencia y videojuegos (versión Extracción de Información) [UCA 05/2015]Ciencia y videojuegos (versión Extracción de Información) [UCA 05/2015]
Ciencia y videojuegos (versión Extracción de Información) [UCA 05/2015]
 
Ciencia y Videojuegos (v2)
Ciencia y Videojuegos (v2)Ciencia y Videojuegos (v2)
Ciencia y Videojuegos (v2)
 
Evolving Evil: Optimizing Flocking Strategies through Genetic Algorithms for ...
Evolving Evil: Optimizing Flocking Strategies through Genetic Algorithms for ...Evolving Evil: Optimizing Flocking Strategies through Genetic Algorithms for ...
Evolving Evil: Optimizing Flocking Strategies through Genetic Algorithms for ...
 
MUSES: A Corporate User-Centric System which Applies Computational Intelligen...
MUSES: A Corporate User-Centric System which Applies Computational Intelligen...MUSES: A Corporate User-Centric System which Applies Computational Intelligen...
MUSES: A Corporate User-Centric System which Applies Computational Intelligen...
 
Publicación de imágenes en la web
Publicación de imágenes en la webPublicación de imágenes en la web
Publicación de imágenes en la web
 
Modelling Human Expert Behaviour in an Unreal Tournament 2004 Bot
Modelling Human Expert Behaviour in an Unreal Tournament 2004 BotModelling Human Expert Behaviour in an Unreal Tournament 2004 Bot
Modelling Human Expert Behaviour in an Unreal Tournament 2004 Bot
 

Recently uploaded

DERIVATION OF MODIFIED BERNOULLI EQUATION WITH VISCOUS EFFECTS AND TERMINAL V...
DERIVATION OF MODIFIED BERNOULLI EQUATION WITH VISCOUS EFFECTS AND TERMINAL V...DERIVATION OF MODIFIED BERNOULLI EQUATION WITH VISCOUS EFFECTS AND TERMINAL V...
DERIVATION OF MODIFIED BERNOULLI EQUATION WITH VISCOUS EFFECTS AND TERMINAL V...
Wasswaderrick3
 
BREEDING METHODS FOR DISEASE RESISTANCE.pptx
BREEDING METHODS FOR DISEASE RESISTANCE.pptxBREEDING METHODS FOR DISEASE RESISTANCE.pptx
BREEDING METHODS FOR DISEASE RESISTANCE.pptx
RASHMI M G
 
DMARDs Pharmacolgy Pharm D 5th Semester.pdf
DMARDs Pharmacolgy Pharm D 5th Semester.pdfDMARDs Pharmacolgy Pharm D 5th Semester.pdf
DMARDs Pharmacolgy Pharm D 5th Semester.pdf
fafyfskhan251kmf
 
The use of Nauplii and metanauplii artemia in aquaculture (brine shrimp).pptx
The use of Nauplii and metanauplii artemia in aquaculture (brine shrimp).pptxThe use of Nauplii and metanauplii artemia in aquaculture (brine shrimp).pptx
The use of Nauplii and metanauplii artemia in aquaculture (brine shrimp).pptx
MAGOTI ERNEST
 
Seminar of U.V. Spectroscopy by SAMIR PANDA
 Seminar of U.V. Spectroscopy by SAMIR PANDA Seminar of U.V. Spectroscopy by SAMIR PANDA
Seminar of U.V. Spectroscopy by SAMIR PANDA
SAMIR PANDA
 
20240520 Planning a Circuit Simulator in JavaScript.pptx
20240520 Planning a Circuit Simulator in JavaScript.pptx20240520 Planning a Circuit Simulator in JavaScript.pptx
20240520 Planning a Circuit Simulator in JavaScript.pptx
Sharon Liu
 
Toxic effects of heavy metals : Lead and Arsenic
Toxic effects of heavy metals : Lead and ArsenicToxic effects of heavy metals : Lead and Arsenic
Toxic effects of heavy metals : Lead and Arsenic
sanjana502982
 
原版制作(carleton毕业证书)卡尔顿大学毕业证硕士文凭原版一模一样
原版制作(carleton毕业证书)卡尔顿大学毕业证硕士文凭原版一模一样原版制作(carleton毕业证书)卡尔顿大学毕业证硕士文凭原版一模一样
原版制作(carleton毕业证书)卡尔顿大学毕业证硕士文凭原版一模一样
yqqaatn0
 
Oedema_types_causes_pathophysiology.pptx
Oedema_types_causes_pathophysiology.pptxOedema_types_causes_pathophysiology.pptx
Oedema_types_causes_pathophysiology.pptx
muralinath2
 
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptxANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
RASHMI M G
 
Nutraceutical market, scope and growth: Herbal drug technology
Nutraceutical market, scope and growth: Herbal drug technologyNutraceutical market, scope and growth: Herbal drug technology
Nutraceutical market, scope and growth: Herbal drug technology
Lokesh Patil
 
NuGOweek 2024 Ghent programme overview flyer
NuGOweek 2024 Ghent programme overview flyerNuGOweek 2024 Ghent programme overview flyer
NuGOweek 2024 Ghent programme overview flyer
pablovgd
 
Comparing Evolved Extractive Text Summary Scores of Bidirectional Encoder Rep...
Comparing Evolved Extractive Text Summary Scores of Bidirectional Encoder Rep...Comparing Evolved Extractive Text Summary Scores of Bidirectional Encoder Rep...
Comparing Evolved Extractive Text Summary Scores of Bidirectional Encoder Rep...
University of Maribor
 
Shallowest Oil Discovery of Turkiye.pptx
Shallowest Oil Discovery of Turkiye.pptxShallowest Oil Discovery of Turkiye.pptx
Shallowest Oil Discovery of Turkiye.pptx
Gokturk Mehmet Dilci
 
Leaf Initiation, Growth and Differentiation.pdf
Leaf Initiation, Growth and Differentiation.pdfLeaf Initiation, Growth and Differentiation.pdf
Leaf Initiation, Growth and Differentiation.pdf
RenuJangid3
 
Mudde & Rovira Kaltwasser. - Populism - a very short introduction [2017].pdf
Mudde & Rovira Kaltwasser. - Populism - a very short introduction [2017].pdfMudde & Rovira Kaltwasser. - Populism - a very short introduction [2017].pdf
Mudde & Rovira Kaltwasser. - Populism - a very short introduction [2017].pdf
frank0071
 
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Ana Luísa Pinho
 
Lateral Ventricles.pdf very easy good diagrams comprehensive
Lateral Ventricles.pdf very easy good diagrams comprehensiveLateral Ventricles.pdf very easy good diagrams comprehensive
Lateral Ventricles.pdf very easy good diagrams comprehensive
silvermistyshot
 
ESR spectroscopy in liquid food and beverages.pptx
ESR spectroscopy in liquid food and beverages.pptxESR spectroscopy in liquid food and beverages.pptx
ESR spectroscopy in liquid food and beverages.pptx
PRIYANKA PATEL
 
Chapter 12 - climate change and the energy crisis
Chapter 12 - climate change and the energy crisisChapter 12 - climate change and the energy crisis
Chapter 12 - climate change and the energy crisis
tonzsalvador2222
 

Recently uploaded (20)

DERIVATION OF MODIFIED BERNOULLI EQUATION WITH VISCOUS EFFECTS AND TERMINAL V...
DERIVATION OF MODIFIED BERNOULLI EQUATION WITH VISCOUS EFFECTS AND TERMINAL V...DERIVATION OF MODIFIED BERNOULLI EQUATION WITH VISCOUS EFFECTS AND TERMINAL V...
DERIVATION OF MODIFIED BERNOULLI EQUATION WITH VISCOUS EFFECTS AND TERMINAL V...
 
BREEDING METHODS FOR DISEASE RESISTANCE.pptx
BREEDING METHODS FOR DISEASE RESISTANCE.pptxBREEDING METHODS FOR DISEASE RESISTANCE.pptx
BREEDING METHODS FOR DISEASE RESISTANCE.pptx
 
DMARDs Pharmacolgy Pharm D 5th Semester.pdf
DMARDs Pharmacolgy Pharm D 5th Semester.pdfDMARDs Pharmacolgy Pharm D 5th Semester.pdf
DMARDs Pharmacolgy Pharm D 5th Semester.pdf
 
The use of Nauplii and metanauplii artemia in aquaculture (brine shrimp).pptx
The use of Nauplii and metanauplii artemia in aquaculture (brine shrimp).pptxThe use of Nauplii and metanauplii artemia in aquaculture (brine shrimp).pptx
The use of Nauplii and metanauplii artemia in aquaculture (brine shrimp).pptx
 
Seminar of U.V. Spectroscopy by SAMIR PANDA
 Seminar of U.V. Spectroscopy by SAMIR PANDA Seminar of U.V. Spectroscopy by SAMIR PANDA
Seminar of U.V. Spectroscopy by SAMIR PANDA
 
20240520 Planning a Circuit Simulator in JavaScript.pptx
20240520 Planning a Circuit Simulator in JavaScript.pptx20240520 Planning a Circuit Simulator in JavaScript.pptx
20240520 Planning a Circuit Simulator in JavaScript.pptx
 
Toxic effects of heavy metals : Lead and Arsenic
Toxic effects of heavy metals : Lead and ArsenicToxic effects of heavy metals : Lead and Arsenic
Toxic effects of heavy metals : Lead and Arsenic
 
原版制作(carleton毕业证书)卡尔顿大学毕业证硕士文凭原版一模一样
原版制作(carleton毕业证书)卡尔顿大学毕业证硕士文凭原版一模一样原版制作(carleton毕业证书)卡尔顿大学毕业证硕士文凭原版一模一样
原版制作(carleton毕业证书)卡尔顿大学毕业证硕士文凭原版一模一样
 
Oedema_types_causes_pathophysiology.pptx
Oedema_types_causes_pathophysiology.pptxOedema_types_causes_pathophysiology.pptx
Oedema_types_causes_pathophysiology.pptx
 
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptxANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
ANAMOLOUS SECONDARY GROWTH IN DICOT ROOTS.pptx
 
Nutraceutical market, scope and growth: Herbal drug technology
Nutraceutical market, scope and growth: Herbal drug technologyNutraceutical market, scope and growth: Herbal drug technology
Nutraceutical market, scope and growth: Herbal drug technology
 
NuGOweek 2024 Ghent programme overview flyer
NuGOweek 2024 Ghent programme overview flyerNuGOweek 2024 Ghent programme overview flyer
NuGOweek 2024 Ghent programme overview flyer
 
Comparing Evolved Extractive Text Summary Scores of Bidirectional Encoder Rep...
Comparing Evolved Extractive Text Summary Scores of Bidirectional Encoder Rep...Comparing Evolved Extractive Text Summary Scores of Bidirectional Encoder Rep...
Comparing Evolved Extractive Text Summary Scores of Bidirectional Encoder Rep...
 
Shallowest Oil Discovery of Turkiye.pptx
Shallowest Oil Discovery of Turkiye.pptxShallowest Oil Discovery of Turkiye.pptx
Shallowest Oil Discovery of Turkiye.pptx
 
Leaf Initiation, Growth and Differentiation.pdf
Leaf Initiation, Growth and Differentiation.pdfLeaf Initiation, Growth and Differentiation.pdf
Leaf Initiation, Growth and Differentiation.pdf
 
Mudde & Rovira Kaltwasser. - Populism - a very short introduction [2017].pdf
Mudde & Rovira Kaltwasser. - Populism - a very short introduction [2017].pdfMudde & Rovira Kaltwasser. - Populism - a very short introduction [2017].pdf
Mudde & Rovira Kaltwasser. - Populism - a very short introduction [2017].pdf
 
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
Deep Behavioral Phenotyping in Systems Neuroscience for Functional Atlasing a...
 
Lateral Ventricles.pdf very easy good diagrams comprehensive
Lateral Ventricles.pdf very easy good diagrams comprehensiveLateral Ventricles.pdf very easy good diagrams comprehensive
Lateral Ventricles.pdf very easy good diagrams comprehensive
 
ESR spectroscopy in liquid food and beverages.pptx
ESR spectroscopy in liquid food and beverages.pptxESR spectroscopy in liquid food and beverages.pptx
ESR spectroscopy in liquid food and beverages.pptx
 
Chapter 12 - climate change and the energy crisis
Chapter 12 - climate change and the energy crisisChapter 12 - climate change and the energy crisis
Chapter 12 - climate change and the energy crisis
 

Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection (CoG 2019)

  • 1. Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection Mohammed Salem (University of Mascara, Algeria) Antonio M. Mora (University of Granada, Spain) J.J. Merelo (University of Granada, Spain) IEEE Conference on Games 2019 20-23 August London, UK
  • 2. Index Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection
  • 3. The Open Racing Car Simulator. Used in Competitions and Testbed for AI Controllers (Driving bots) Inputs/Outputs: – Sensors: Car/Track distances – Actuators: Steer, Break, Accelerator, Gear http://torcs.sourceforge.net/ TORCS Racing bot
  • 6. Target Speed Fuzzy Subcontroller Racing bot Mamdani Fuzzy System with 3 Trapezoidal Membership Functions (Low, Medium, High).
  • 7. Target Speed Fuzzy Subcontroller Racing bot IF Front = High THEN TargetSpeed is 280 IF Front = Medium THEN TargetSpeed is 240 IF Front = Low AND M5 = High THEN TargetSpeed is 220 IF Front = Low AND M5 = Medium THEN TargetSpeed is 180 IF Front = Low AND M5 = Low AND M10 = High THEN TargetSpeed is 120 IF Front = Low AND M5 = Low AND M10 = Medium THEN TargetSpeed is 60 IF Front = Low AND M5 = Low AND M10 = Low THEN TargetSpeed is 30 IF Front = MAXDISTSPEED OR M5 = MAXDISTSPEED OR M10 = MAXDISTSPEED THEN TargetSpeed = 300 Uses a Set of Fuzzy Rules:
  • 8. Steer Control Fuzzy Subcontroller Racing bot Mamdani Fuzzy System with 3 Trapezoidal Membership Functions. Same inputs (FRONT, M5 and M10). Set of Fuzzy Rules: IF Front = High THEN steer = 0 IF Front = Medium AND M10 = High THEN steer = 0.25 IF Front = Medium AND M10 = Medium AND M5 = Medium THEN steer = 0.25 IF Front = Medium AND M10 = Medium AND M5 = Low THEN steer = 0.5 IF Front = Low AND M10 = High THEN steer = 0.5 IF Front = Low AND M10 = Medium AND M5 = Medium THEN steer = 0.75 IF Front = Low AND M10 = Medium AND M5 = Low THEN steer = 0.75
  • 9. Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection Index
  • 11. Genetic Algorithm evolution Population: set of controllers defined each one by a vector which determines the shape of the membership functions. Evaluation: every candidate controller must be evaluated in the environment (TORCS races). A value must be assigned to its performance => Fitness Function.
  • 12. Genetic Algorithm evolution Fitness Function: - Maximize average speed on the whole circuit while minimizing damage. - Focuses on objectives of a (human) driver during a race, not just on winning or losing => human-like.
  • 14. Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection Index
  • 15. Challenges uncertainty Reliable and effective evaluation process:  Good individuals <=> Competitive controllers.  Selection of the best. Variable results for the same controller:  Weather conditions.  Track status.  Non-deterministic behavior of rivals.
  • 16. Solutions uncertainty Fitness definition and evaluation:  Considering the main factors (just speed and damage).  Several races.  Several laps.  No rivals.  Track combining difficult and easy parts. Alpine 2 Track
  • 17. Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection Index
  • 18. Enhanced Exploration Blend Crossover Operator (BLX-a):  Generates different descendants with random genes.  The differences (depend on a value).  If a > 0.5 => higher exploration  If a < 0.5 => higher exploitation Two approaches:  a constant  a variable
  • 19. Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection Index
  • 20. Pole position selection Or race-based selection:  Individuals arranged in groups of 10.  They compete in several races of several laps (same car).  They obtain a F1-based score depending on their final position in every race.  At the end, those 5 with the best global score are selected as parents. Advantages:  Less sensitive to uncertainty (fitness-less selection)  Selected individuals will be almost surely the best.
  • 21. Beating uncertainty in racing bot evolution through enhanced exploration and pole position selection Index
  • 22. Experimental Setup beating GA parameters:  Generations = 50  Population size = 60  Crossover rate = 0.85  Mutation rate = 0.09 Genetic operators:  Random initialization  Race-based selection [every 5 generations] Tournament selection [the rest]  Non-uniform mutation  BLX-a crossover - CONSTANT => a = 0.5 - VARIABLE => a = 1-(generation/TOTAL_GENS) Evaluation:  Alpine 2 Track, 20 laps, Solo race, car1-tbr1 (moderate performance)
  • 23. Selection of the best beating At the end of GA run:  10 best individuals (according to fitness).  5 races.  5 laps.  Alpine 2 Track.  Random starting grid.  F1-based score: 1st - 25 points, 2nd - 18, 3rd - 15, 4 - 12, 5 - 10, 6 - 8, 7 - 6, 8 - 4, 9 - 2, 10 – 1. At the end of 10 runs:  The 10 best individuals (selected before).  5 races/5 laps in Alpine 2.  Random starting grid.
  • 24. Results beating Facts:  New approaches improve previous controllers.  GFC-VA clearly outperforms all the rest.  Excellent performance even in an unknown track (E-Track 5). Controllers:  GFC -> our previous best controller (AVS Fitness).  GFC-RS -> our previous controller with race-based selection.  GFC-FA -> new controller with race-based selection and BLX-a with constant value.  GFC-VA -> new controller with race-based selection and BLX-a with variable value.
  • 25. Results beating Facts:  New approaches improve previous controllers.  GFC-VA clearly outperforms all the rest.  Excellent performance even in an unknown track (E-Track 5). Controllers:  GFC -> our previous best controller (AVS Fitness).  GFC-RS -> our previous controller with race-based selection.  GFC-FA -> new controller with race-based selection and BLX-a with constant value.  GFC-VA -> new controller with race-based selection and BLX-a with variable value.
  • 26. PSRI Controller beating New Challenger: PSRI  Controller created by Pérez- Liébana, Recio, Sáez and Isasi.  4th in Simulated Car Racing Competition 2009.  Finite State Machine using fuzzy rules for transitions.  Optimized by means of NSGA- II.  J48 classifier for prediction of parts of the track.
  • 27. Results beating Facts:  Both GFC-FA and GFC-VA improve all PSRI results.  PSRI is able to win one of the races on each track.  GFC-VA is the best of the three (wins 4 and 3 races). 5 races / 20 laps per track. Average results.
  • 28. We have Future Work:  Implement completely Race-based selection (in all the generations).  Optimize the output of the fuzzy rules.  Optimize the set of fuzzy rules (Genetic Programming). Beaten uncertainty in racing bot evolution through enhanced exploration and pole position selection conclusions