Overview / Usage

Nowadays I am trying to build a robotic arm that will have two camera. One for object detection and the other is for detection of distance of objects. In the final case, i will put lots of object in front of the arm randomly and i say that (or write as typing to arm system) "Grasp a plastic bottle" the arm will detect are there or not a plastic bottle among the other objects. After that if there was, the dept sensor detect position of the bottle. After detection arm start to move for grasping it , but i will not hand coded its movement before. So it will find the way how it will reach the object with reinforcement learning technique
.
However to build such a system I have to have 4 thing

1. Knowledge of Reinforcement learning and Neural Net for object detecting .
2. Robotic arm body and motors raspberry and other things.
   3 . Depth camera

Actually ,
I have already have knowledge on reinforcement learning and neural net and i am continue to improve it
I have already 3d printer that it could be print robotic arm, and also i have raspberry pi 2 with camera but i could not know how could i handle Neural net calculation on the pi? I guess i will need much more power ...
And final think i havent got any depth camera , when i search online it seams much expensive according my personal bugged .

Lets start and see what will be happen ....