MCB419 logo  


Red ball 1

Simple bouncing ball

Red ball 2

Bouncing ball with user interface.

Red ball 3

Single ball, object-oriented design (no constructor).

Multi balls 1

Multiple balls using object-oriented design (with constructor).

Multi balls 2

Change representation from (vx, vy) to (speed, heading), which is more natural for motor control.

Wander 1

Wander behavior.

Wander 2

Wander behavior; add trails and wanderNoise slider.

Twitchy bugs

Wandering bugs with too much caffeine (uses uniform random noise).

Perlin bugs

Wandering bugs with smoother motion using Perlin noise; also nice watery Perlin-noise background.

Week 3 homework

Exploring the effects of wander noise on foraging efficiency.

Week 4 homework

Different kinesis strategies influence the spatial distribution of agents.

Week 5 homework

Braitenberg vehicles

Week 6 homework

Using GAs to evolve Braitenberg controllers

Week 7 homework

Evolving a C. elegans inspired controller

Week 7 homework

A Braitenberg controller using integrate-and-fire neurons

Week 7 homework

Solving the XOR problem with Threshold Logic Units

Week 8 homework

Action selection / Finite-state machines / Area-restricted search

Week 9 homework

Lamprey place-code to rate-code conversion; integrate-and-fire neurons

Week 11 homework

Associative learning, estimating rewards, action policies

Week 12 homework

Reinforcement learning in a grid world; Q-learning with tables

Week 13 homework

Reinforcement learning with neural networks

Mark E. Nelson, University of Illinois at Urbana-Champaign, 2005-2016.