Material

• Jan's master thesis with an introduction to Kalman filtering (see chapter 2.1). There you can also see how a Kalman filter can be used for map creation and localisation with a camera.
• Simulator dumps. Jan's ball-on-a-tiltable-board-observed-by-a-silicon-retina simulator was used to create two dumps (with and without noise). Each dump consists of two files: retina.dump and world.dump. Each line in the retina dump consists of timestamp, x, y, and direction of the brightness change (1=brighter,-1=darker). Each line in the world.dump consists of timestamp, x position, y position, x speed, y speed, x acceleration, and y acceleration of the ball.
• Jan's simulator (runs on linux only). You can get it via git: git clone http://janeway.inf.tu-dresden.de/~jan/retina. After checkout, change into the retina directory, type cmake ., and realise that you are missing some fancy packages. On Ubuntu, these are: libboost-1.42-all-dev, freeglut3-dev. If you got them, running cmake . should not complain anymore (assuming that you have some elementary C++ development packages already). Type make and if everything compiled fine, run the simulator via ./retina.