Now, there's some funny stuff over here.

You can look at these numbers carefully. It ranks the corner points more into the interior.

If U equals 1, all the values are cyclically shifted to the right by 1.

So while there's normalization over here,

If U equals 3, they are cyclically shifted to the right by 3.

Let's assume we have a 0.5 in this cell and a 0.5 in this cell.

That's because you go backwards in time.

This is a truncation, indicating we have a cyclic array, and we do the same with j let me scroll very carefully.

We just went off the y value.

I want to do it in a way that is cyclic, so that the last point connects smoothly to the initial point over here.

With 0.1 chance we don't move at all, and with 0.1 chance again we skip and move 2 steps.

We told you the robot lives in a world of size 100 x 100, and what this means if this robot falls off 1 end, it disappears on the other.

I want you to program a function that returns the new distribution Q after the move where if U equals zero, Q is the same as p.

In fact, the way we're going to program is we will assume the world is cyclic, so if it drops off the right most cell it finds itself in the left most cell.

I go through all the i's as opposed to leaving out the first and last node, and because we are cyclic, I now have to add the model of division by the length of the path in this smoothing parameter.

We're just going to add lines of code that achieve this over here, and when you do this, please make sure that your update understands the fact that these indices are cyclic.

1 9, 1 3, 1 3, 1 9, and 1 9 and even though we don't know where the robot is, the robot moves, and it moves to the right.

Then we have a number of motion commands 0, 0 is no move, 0, 1 means you move to the right, 0, 1 means move left, 1, 0 makes you move down, not up, and 1, 0 makes you move up.

For a node like this with coordinates xi and yi, we add to it with a very small constant gamma in fact, it will be half our weight smoothing constant in a minute 2 times the previous guy and of course that's cyclic, so you have to make sure this is really cyclic minus the guy 2 steps away and minus our node.

What's going to happen to my robot if it never senses but executes the action of going 1 to the right on our little cyclic environment forever.?

In class, we localized the robot in a 1D world with a number of grid cells where each grid cell could have a different color, and the measurement vector was a sequence of observations.