The only difference I see is that you have both True, and False level triggering enabled but that still works in my file. Can you try uploading your file again? I think thats the only way I’ll be able to help any further, or you could try one of the other methods, users hae posted in this thread.
Ok, found the problem. Your grid is much larger than the grid in my example. My grid had 1bu x 1bu grid spaces, and yours is 2bu x 2bu. You just have to adjust your file to that by doubling the range your rays. I also noticed that you need to change the direction of some of your rays as well. Then It should work fine.
Here are the steps I took to get this to work, though @kevincorrigan just beat me to it:
-Switch from Blender Render to Blender Game 
-Change the ‘collision’ sensor for sensing Win to a near sensor, set to a range of 0.01
-Check ‘actor’ under the physics tab for both the robot and the win cube.
This gets the success state to function.
-In Target’s logic, changed the length of all the Ray sensors to 1.0.
-Fixed the axis of the Ray sensors for forward and right (they need to be negative)
This fixes the collision issue, and seems to work quite well.
A peice of advice: The cubes making up the floor and walls of that scene must be vastly optimised. There’s no reason those cubes need to be 1000 polys each
Thank you all so much for helping me. The game is finally working!