A single-point-positioning solution at one receive epoch.

position is the converged ITRF/IGS ECEF position in meters. geodetic is the same point as %{lat_rad, lon_rad, height_m} when geodetic output was requested (the default), otherwise nil. rx_clock_s is the reference-system receiver clock bias in seconds; system_clocks_s is a map of GNSS letter (e.g. "G", "E") to that system's absolute receiver clock in seconds (a single entry for a one-system solve, one per constellation for a mixed solve). These are per-system clocks, not biases: the inter-system bias of a system is its clock minus the reference system's (rx_clock_s). dop carries the dilution-of-precision scalars for any full-rank geometry — a single-system solve uses the bit-exact four-state cofactor, a multi-system solve a general inverse with one clock column per constellation — and is nil only when the geometry is rank-deficient. residuals_m are the post-fit pseudorange residuals in meters, in used_sats order. used_sats are the contributing satellite id strings (e.g. "G01"); rejected_sats pairs each excluded satellite id with its reason atom (:no_ephemeris or :low_elevation). metadata reports solver iterations, convergence, the corrections applied, and the geometry redundancy: used_count, the distinct systems, the redundancy (degrees of freedom, used_count - (3 + systems)), and raim_checkable? (redundancy >= 1). An exactly determined fix has redundancy < 1, forcing the residuals near zero and leaving the fix unverifiable by RAIM. When the opt-in :huber reweighting runs, metadata also carries :huber with the outer_iterations count and the final_scale_m (the last MAD robust scale in meters); the key is absent on the default static path.