Hi Mirzel,

Nice animation!  I think it could work in theory but so far practice has not justified theory with permanent magnet setups...  unfortunately.

I attached a picture from one of your videos I grabbed it at a moment when opposite poles are just happen to face each other ( I drew small lines to show where).  In this moment of rotation there must be a huge sticky/dead point against rotation and I understand your aim to use 2 more similar arrangements connected to the same shaft to help go through on these dead points but question is how much useful torque remains (if remains) at all?  And using stronger and stronger magnets to compensate and increase torque may help (once assuming you have reached self rotatation...)  till you reach the best magnets available as the final limit? 
Another problem maybe is that stronger magnets have certain volumes and weights and you place them excentrically on the wheels which have supposedly ball bearing shafts so the wear or mechanical load on the bearings can be huge after a certain time.

But all these would not count if your setup is able to run on its own...  and the only way to learn about it is to make a prototype.  I wish you good luck in this.

rgds,  Gyula