[Payload]Trade-off CoM and payload movements
Closed this issue · 4 comments
This issue gathers progress on the goal of formulating a robust and as generic as possible cost function to regulate the relative CoM and payload moves. The goal is that the cost function provides the user the ability to determine how will the necessary mass translation be achieved, by moving the CoM, the payload (that is the hands), both of them and which one more?
We define the cos_fraction = cost_CoM / cost_hands
, with cost_CoM = 1e3 = constant
- Plots with
cost_fraction = 1 (cost_hands = cost_CoM = 1e3)
- Plots with
cost_fraction = 0.001 (cost_hands = 1)
It is obvious that the hands relative movement, velocity and acceleration are much lower for the case of equal cost (cost_fraction = 1). Moreover the hands do not finish in the default position (this can be corrected with a high final cost). CoM movement is affected but not so much
We can see that by modifying the cost_fraction the hands' movement is regulated, however the CoM movement is not affected a lot. Thus this strategy does not trade off the CoM and hands moves.