PositionVelocityTime v7.0.0

7.0.0 (2026-10-16)

feat(velocity)!: share the pseudorange-rate model as calc_pseudorange_rate and fit it in one step (f88accd)
fix(velocity)!: model the range rate as the derivative of the Earth-rotation-corrected pseudorange (72524e9)
refactor!: share the Earth-rotation correction, rank-check the velocity solve, spell pseudoranges (8fee0a4)

calc_line_of_sight is no longer part of the documented API; use calc_range_rate for the range-rate model, or the negated position columns of calc_H/calc_H! for the line of sight. The undocumented rotate_by_earth_rotation is removed. calc_user_velocity_and_clock_drift[!] return NaN for the state and every residual, instead of throwing, when the velocity design is not of full rank to within the tolerance calc_DOP applies.

Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com

calc_user_velocity_and_clock_drift! takes a design-row scratch vector (Vector{SVector{4,Float64}}) and the solved user position in place of the position design H, and an optional start state, calc_user_velocity_and_clock_drift!(rate_residuals, design, measurements, user_pos, start = zero(SVector{4,Float64})). calc_user_velocity_and_clock_drift(measurements, user_pos, start = zero(SVector{4,Float64})) likewise takes the user position in place of H. The design comes from the pseudorange-rate model, calc_pseudorange_rate, which the solve fits in one Gauss-Newton step.

Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com

calc_user_velocity_and_clock_drift[!] take the solved user position (ECEF, m) as a last argument, after H: calc_user_velocity_and_clock_drift!(rate_residuals, measurements, H, user_pos). The range-rate model needs it for the travel time, its rate and the Earth-rotation-corrected satellite.

Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com

Merged pull requests:

fix(velocity)!: model the range rate as the derivative of the Earth-rotation-corrected pseudorange (#94) (@giove-a)