suppress medical pump generated flow disturbances with RL.

Pumps in medical applications require a steady and controllable flow. We designed a medical pump consisting of valves and pistons, but due to the limited length of the pistons and the fluid displacement by the valves, the dynamics of this pump are discontinuous and non-linear. With traditional control algorithms, it is difficult to design a feedback controller that controls flow accurately and suppresses disturbances sufficiently.

10 times better.

We successfully implemented a reinforcement learning-based feedback controller that suppresses the disturbances caused by variations in the pressure of the source as well as variations caused due to production tolerances of the pump itself.

The trained controller suppresses the disturbance better than the traditional controller. This results in less peaking of the flow while the pump is in use during a medical procedure.