Visualisation of Trajectory Replanning

About a bazillion years ago In 2008, as part of an assignment for my Masters degree, I wrote a visualiser for Trajectory Replanning in Processing. Here it is, converted to Processing.js:

Trajectory Replanning?!

Yes:

  1. I start at the green dot.

  2. I want to get to the blue dot.

  3. I don’t have perfect information about the maze. I learn as I explore.

  4. I’ll use A* to figure out how to get to the blue dot.

  5. Hey, I moved a bit and discovered a wall I didn’t know about! My plan is now useless.

So, the assignment was to evaluate 3 different methods for coping with the change in the agent’s knowledge of the world that happend at step 5 above:

  1. Repeated Forward A*

  2. Repeated Backward A*

  3. Adaptive A*

It’s late, and the purpose of this was to just get the visualisation online, so I’m not going to explain the various algorithms. Code is on GitHub, and you can check out my assignment report, if you like (PDF warning…).