Simulator programmed in Python with PyGame. You can start by running the file:
Where you can define the iterations, the population size, the selected percentage, the error range and the mutation percentage.
The Main purpose of the code is to Create a neural network in combination of an evolutional algorithm for a robot simulator robot.
The evolutionary alogirthm implemented, trains the robots in several rooms at each generation and uses the average fitness as robot fitness to select. It can use one or two points crossover, mutation and elitism.
- Mobile robot simulator
- ANN as controller
- EA for evolve weights of ANN
- Differnet room designs
- Used ANN as controller
- Used two layers with recurrent nodes
- Used feedback to create memory
- Play with Dt (depend on time step)
- Input: 12 infrared distance sensors (30°distance)
- Output: two outputs – each controls speed of one wheel
- Collision-free
- Fix time for each experiment
- Simulate dust, used removed dust as fitness
The program is in Python
In order to use the code you need to install the following packages
import numpy as np
import math
import pygame
import time
import random
import matplotlib.pyplot as pltElena Kane
Nikolaos Ntantis
Ioannis Montesantos