With reference to #161
The visualization is in accordance with the partially observable maze problem referenced in chapter 4. Please go through it to completely understand what I am trying to do here.
There are 3 parts to it (as of now).
1) We first introduce the concept of percepts in a maze where the robot can only see if there is a wall beside it. The user is given a specific percept and he/she needs to click on all the cells of the maze where the robot may get that percept.
2) We show an abstract solution to the above problem by allowing the user to hover over any cell and see its percept. Along with it, it also highlights all other cells which can have the same percept. This is to introduce the concept of belief states (states where the robot believes it can be)
3) We start with an actual robot (hidden from the user) and its corresponding belief states. User needs to pinpoint the location of the robot by moving around the robot. The user can move the robots around and see how some of them gets eliminated when the percepts do not match with the actual percept.
The actual problem in the book has only one action "move" where the robot will move in some random direction. This would be easier to explain after the first 3 diagrams.
The visualization is still incomplete and any review or criticism is welcomed.
I will continue working on it if is still useful to aima-javascript.
(Work in progress. Do not merge)
With reference to #161 The visualization is in accordance with the partially observable maze problem referenced in chapter 4. Please go through it to completely understand what I am trying to do here. There are 3 parts to it (as of now). 1) We first introduce the concept of percepts in a maze where the robot can only see if there is a wall beside it. The user is given a specific percept and he/she needs to click on all the cells of the maze where the robot may get that percept. 2) We show an abstract solution to the above problem by allowing the user to hover over any cell and see its percept. Along with it, it also highlights all other cells which can have the same percept. This is to introduce the concept of belief states (states where the robot believes it can be) 3) We start with an actual robot (hidden from the user) and its corresponding belief states. User needs to pinpoint the location of the robot by moving around the robot. The user can move the robots around and see how some of them gets eliminated when the percepts do not match with the actual percept.
The actual problem in the book has only one action "move" where the robot will move in some random direction. This would be easier to explain after the first 3 diagrams. The visualization is still incomplete and any review or criticism is welcomed. I will continue working on it if is still useful to aima-javascript. (Work in progress. Do not merge)