At last, I have completed my artificial intelligence that learns similar to how IBM Watson learns. It is a very complicated system that required a lot of testing and… brain warping to finish.
I was curious if any other teams had a similar type of program. I know my program has many bugs and issues that I should work out for the 2017-2018 Game.
I won’t be sharing my code for the sake of keeping it limited. I will share some pseudo code though.
Here is my team’s autonomous robot’s code: https://vexforum.com/t/wpi1-reveal/41037/1
Why don’t you share some videos of the robot playing the game and your pseudo-code.
Crud. I’ll type some pseudo code soon. Robot isn’t operational at the moment. Oops.
Aditya…
What r u workin on bro…
To what extent does the program learn? This sounds like a very interesting program. How did you give the program input on its performance?
I could not get my hands on Lidar sensors. As soon as I do, I will integrate those into the program in the neatest manner that I can do.
This basically takes object locations from a file that I give in my tile grid system. It then can find the fastest route and naturally moves to the destination; rather than a machine moving type system. It learns the methods to slow down and speed up based on the route completition percentage. Through 100s of tests, the robot is then able to figure out how to drive to an object. Sadly, file pulling is above me at this point in time so every time I upload a program to the robot - like a baby - it will know nothing other than the machine moving, so I would have to reteach the natural movement.
Attaboi- Thats my 2017-2020 programmer right there