Permalink
Switch branches/tags
Nothing to show
Commits on Nov 26, 2011
  1. Typos in comments.

    Martin J. Laubach committed Nov 26, 2011
Commits on Nov 24, 2011
  1. Oops, wrong comparison for check.

    Martin J. Laubach committed Nov 24, 2011
Commits on Nov 23, 2011
  1. Drawing lots of particles is painfully slow. Add option to only draw …

    Martin J. Laubach committed Nov 23, 2011
    …every n-th particle so the observer doesn't die of old age.
  2. Make maze a bit more interesting

    Martin J. Laubach committed Nov 23, 2011
  3. Optimise picking a bit by not including particles with w=0. If no eli…

    Martin J. Laubach committed Nov 23, 2011
    …gible particle is left, then we reinitialise to a random state.
  4. Nicer shape for directed particles

    Martin J. Laubach committed Nov 23, 2011
Commits on Nov 21, 2011
  1. Add option for having a robot without compass, ie. it has no idea whi…

    Martin J. Laubach committed Nov 21, 2011
    …ch direction it is heading.
  2. Prepare for handling directed particles

    Martin J. Laubach committed Nov 21, 2011
  3. Add triangle shape for directed belief states

    Martin J. Laubach committed Nov 21, 2011
Commits on Nov 20, 2011
  1. Simplify means calculation

    Martin J. Laubach committed Nov 20, 2011
  2. Refactor

    Martin J. Laubach committed Nov 20, 2011
  3. Compute center weighted by particle weights

    Martin J. Laubach committed Nov 20, 2011
  4. Tweak gaussian sigma a bit narrower.

    Martin J. Laubach committed Nov 20, 2011
  5. Reduce number of draws even further

    Martin J. Laubach committed Nov 20, 2011
  6. When all particles are improbably, the picker would crash. Avoid that…

    Martin J. Laubach committed Nov 20, 2011
    … by adding random particles in that special case.
  7. Clean up comments a bit

    Martin J. Laubach committed Nov 20, 2011
  8. Use real distance instead of distance^2

    Martin J. Laubach committed Nov 20, 2011
  9. Optimise painfully slow drawing a bit: avoid drawing particles at the…

    Martin J. Laubach committed Nov 20, 2011
    … same place on screen several times.
  10. Update .gitignore

    Martin J. Laubach committed Nov 20, 2011
Commits on Nov 19, 2011
  1. Move mean point computing to own helper function

    Martin J. Laubach committed Nov 19, 2011
  2. Remove dead code

    Martin J. Laubach committed Nov 19, 2011
  3. Speed up robot, let it change direction less often, rearrange maze to…

    Martin J. Laubach committed Nov 19, 2011
    … be more easily identifiable
  4. Add option to only draw every n-th step, so experimenting with large …

    Martin J. Laubach committed Nov 19, 2011
    …particle counts is less painful
Commits on Nov 18, 2011
  1. Add a mean point display to show how good the robot's position has be…

    Martin J. Laubach committed Nov 18, 2011
    …en determined.
  2. Clarify and remove dead code.

    Martin J. Laubach committed Nov 18, 2011
  3. Add comments.

    Martin J. Laubach committed Nov 18, 2011
  4. add .gitignore

    Martin J. Laubach committed Nov 18, 2011
  5. Always run the robot at a fixed speed in a random direction instead o…

    Martin J. Laubach committed Nov 18, 2011
    …f random speed in random direction, we don't gain anything from slowly pacing around and it actually hurts.
  6. Use jtosey's optimised weighted picking method.

    Martin J. Laubach committed Nov 18, 2011
  7. Use jtosey's distance constants

    Martin J. Laubach committed Nov 18, 2011
Commits on Nov 17, 2011
  1. Move moving after reshuffling to be in line with algo, tweak paramete…

    Martin J. Laubach committed Nov 17, 2011
    …rs for better demo effect.
  2. Rename m -> world

    Martin J. Laubach committed Nov 17, 2011