[Kalman] Added support for control inputs in Kalman class

Before, Kalman class did not support control inputs in the estimation step of the algorithm. This feature is required in the algorithm that estimates the mass of the rocket, so I added generic support for M inputs, which required a new matrix in KalmanConfig and a new template argument. I made sure that the change does not break ADA and NAS algorithms, and fixed all usages of Kalman class with the new template argument.