Function correct_kinematic_state 

pub fn correct_kinematic_state(
    source: &dyn ObservableEphemerisSource,
    epoch: &FloatEpoch,
    state: &mut KinematicState,
    covariance_m2: &mut Vec<Vec<f64>>,
    config: &KinematicConfig,
) -> Result<KinematicUpdateSummary, KinematicSolveError>

Correct one kinematic PPP EKF state with a single epoch of code/phase rows.

This uses the same shared PPP model row builder as the static float solver, then applies an EKF measurement update with diagonal measurement covariance derived from those rows’ inverse-sigma weights.