Motocross And Artificial Neural Networks

1
Motocross And Artificial Neural Networks
Benoit Chaperot, Colin Fyfe, School of
Computing, The University of Paisley, Paisley,
PA1 2BE, SCOTLAND.
2
Why use Artificial Neural Network

• Riding a motorbike involves behaviours which are
  difficult to express as a set of procedural rules
• ANN expected to behave in a human or animal-like
  manner
• Capable of extrapolating when presented with new
  and different sets of inputs
• Capable of evolving

3
The ANN

• The inputs
• Bike position, orientation and velocity relative
  to the track
• Terrain height information
• Track path information

The outputs Accelerate, brake Turn left,
right Lean forward, backward
ANN
4
Two forms of training

• Evolutionary algorithms Training considered as
  an optimisation to be performed using genetic
  algorithms
• Back propagation algorithm ANNs are trained
  using training data made from a recording of the
  game being played by a good human player

5
Evolutionary Algorithm

• Training considered as optimisation
• ANNs initialised with random weights
• 80 ANNs per generation
• Each ANN evaluated for 60 seconds or less using a
  score function
• Fittest ANNs are given more chance to reproduce
• The whole population converges to a satisfactory
  solutions to the problem after approximately 100
  generations
• May be difficult to find a good evaluation
  function

6
Fitness Function

• Bonus for passing through a way point
• Bonus/penalty (i.e. normally negative) for
  missing a way point
• Bonus/penalty (i.e. normally negative) for
  crashing
• Bonus/penalty (i.e. normally negative) for every
  meter away from the centre of the next way point

7
Back propagation algorithm

• Training data made by the main author playing the
  game on many different tracks.
• Each sample of training data contains a situation
  (bike position, orientation on a track) and the
  main authors solution to the situation
• Training data composed of approximately 120000
  samples
• Good solution to the problem after only 20000
  iterations

8
Results

• ANNs learn and perform like a human intelligence
• Average lap time
• Good human player 2 min 10 sec
• ANN trained using GA 2 min 50 sec
• ANN trained using BP 2 min 20 sec
• ANN trained using GA slow, but better than one
  trained using BP at adapting to new situations
• ANN trained using BP performance improved using
  bagging and boosting techniques

9
Video

• Videod.avi

10
Bagging and Boosting

• Create N bags by randomly sampling from data set
  with replacement
• P(datum in bag) 0.67
• Machines trained on bags separately
• Results combined
• Boosting puts more emphasis on data which
  machines trained on early bags find difficult
• Anti-boosting seems to work well !

11
Future work

• Use evolutionary algorithms with a much larger
  population size and number of generations
• Further use of bagging techniques
• Use of swarm intelligence techniques