Take the previous loop's state vector, and use it as the initial condition to estimate the current state vector using the IMU's linear acceleration measurements.

Velocity updates should follow the form v_i = v_i_prev + a_i*dt where v_i is the updated i-component of the velocity vector, v_i_prev is the previous loop's i-component of the velocity vector, a_i is the i-component of the IMU's linear acceleration measurement, and dt is the time difference between the current loop and the previous loop.

Position updates should follow the form r_i = r_i_prev + v_i_prev*dt + 0.5*a_i*dt^2.

Attitude Euler angles should be a straightforward overwrite of the current values with the latest IMU estimate, since the IMU handles absolute attitude estimation itself.

Note: We do not need to update the mass estimate throughout the flight.