Complex behavior without neural networks and training. The philosophical basis


The task of motion control is interesting in itself. How to make a mechanical doll solve problems in three-dimensional space? Even if simple. Go to the table, take the item, move it, put it in the basket. Do not fall down the road. Do not bump anywhere.


Boston Dynamics robots are highly inspired. The things they do are amazing. How to get something similar at home? How do you approach this task and how feasible is it?


Hierarchical management


, . , , , . .


, . . , , . . , .


" " : " , , ". , , , , . — . , , , , , .


. , . . , , . , .


, . , . . , , . ?


, . , . . , . , . . , . . , , ( ), , , .


, " ", " ", " " . . , . .


?


. . .
?
? . .
? .
? . ? , . , .


, . .
?
? .
? .


. .
? ? ? , .
? .
? , .
, .
? - , .


. .
, , .
. .
, .
? .


. . .
, , , , .


. , : , . ? . , - . !


, ?


. . , , , , ? , , , ? : . — , . . . — , . , . . , . , .


-. , . . . , , — . , . , .



, , . . . , , . , , . , , , . — . , , . , , . .


. — . , , , . , . . , .


" ( )" .., () , . . , , . , , , , . , . , , , . , , , . , , .



, , . , . . , , . . , . , - . - - . .


Next time I will tell you how the controllers of the joints and limbs can be arranged.
There will be a lot of linear algebra and a little theory of automatic control. That's all for now.



All Articles