kshana 0.21.0

Open, reproducible PNT-resilience simulator with quantum-sensor performance models
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
// SPDX-License-Identifier: AGPL-3.0-only
//! Measurement-domain GNSS simulation: pseudoranges with ionospheric and
//! tropospheric delays.
//!
//! The geometric constellation pack reports visibility and dilution of precision;
//! this module goes one layer deeper and synthesises the actual **pseudorange**
//! each visible satellite would produce, so a downstream solver or RAIM can work
//! in the measurement domain. The modelled pseudorange is
//!
//! ```text
//!   ρ = r_geom + c·δt_rx − c·δt_sv + I + T + ε
//! ```
//!
//! * `r_geom` — the true geometric range (ECEF), the only term in the purely
//!   geometric pack.
//! * `c·δt_rx`, `c·δt_sv` — the receiver and satellite clock offsets (m).
//! * `I` — the **ionospheric** group delay from the Klobuchar single-frequency
//!   model (IS-GPS-200 §20.3.3.5.2.5).
//! * `T` — the **tropospheric** delay: a Saastamoinen zenith delay (Davis et al.
//!   1985; Groves §9.4) projected with the Niell (1996) mapping function.
//! * `ε` — zero-mean thermal-noise plus a deterministic multipath term.
//!
//! Each model is documented with its reference and unit-tested against hand
//! computation; a zero-noise pseudorange reproduces `r_geom + clocks + I + T` to
//! sub-millimetre.

use crate::frames::{geodetic_to_ecef, is_visible, look_angles, teme_to_ecef, Geodetic, Vec3};
use crate::scenario::TimeCfg;
use crate::walker::{walker_epoch_jd, WalkerSgp4};
use rand::SeedableRng;
use rand_chacha::ChaCha8Rng;
use rand_distr::{Distribution, Normal};
use serde::{Deserialize, Serialize};
use sha2::{Digest, Sha256};
use std::f64::consts::PI;

/// Speed of light (m/s).
pub const C_M_PER_S: f64 = 299_792_458.0;
/// GPS L1 carrier frequency (Hz).
pub const L1_HZ: f64 = 1_575_420_000.0;

// ----------------------------------------------------------------------------
// Ionosphere — Klobuchar single-frequency model (IS-GPS-200 §20.3.3.5.2.5).
// All angles in the algorithm are in **semicircles** (1 sc = 180° = π rad); time
// is GPS seconds of day at the sub-ionospheric point.
// ----------------------------------------------------------------------------

/// The eight broadcast Klobuchar coefficients (αₙ in s/semicircleⁿ, βₙ in
/// s/semicircleⁿ) from subframe 4, page 18.
#[derive(Clone, Copy, Debug)]
pub struct KlobucharCoeffs {
    pub alpha: [f64; 4],
    pub beta: [f64; 4],
}

impl Default for KlobucharCoeffs {
    /// A representative mid-activity broadcast set (IS-GPS-200 worked example).
    fn default() -> Self {
        Self {
            alpha: [3.82e-8, 1.49e-8, -1.79e-7, 0.0],
            beta: [1.43e5, 0.0, -3.28e5, 1.13e5],
        }
    }
}

/// Klobuchar L1 ionospheric group delay (metres) for a satellite at geodetic
/// `lat_rad` / `lon_rad`, elevation `el_rad`, azimuth `az_rad`, and `gps_sod`
/// (GPS seconds of day). Returns the slant delay added to the pseudorange.
pub fn klobuchar_delay_m(
    coeffs: &KlobucharCoeffs,
    lat_rad: f64,
    lon_rad: f64,
    el_rad: f64,
    az_rad: f64,
    gps_sod: f64,
) -> f64 {
    // Work in semicircles.
    let phi_u = lat_rad / PI;
    let lambda_u = lon_rad / PI;
    let e = el_rad / PI;
    let a = az_rad; // azimuth used as radians in cos/sin below

    // Earth-centred angle (semicircles).
    let psi = 0.0137 / (e + 0.11) - 0.022;
    // Sub-ionospheric latitude (semicircles), clamped to ±0.416.
    let mut phi_i = phi_u + psi * a.cos();
    phi_i = phi_i.clamp(-0.416, 0.416);
    // Sub-ionospheric longitude (semicircles).
    let lambda_i = lambda_u + psi * a.sin() / (phi_i * PI).cos();
    // Geomagnetic latitude (semicircles).
    let phi_m = phi_i + 0.064 * ((lambda_i - 1.617) * PI).cos();

    // Local time (s), wrapped to [0, 86400).
    let mut t = 43_200.0 * lambda_i + gps_sod;
    t = t.rem_euclid(86_400.0);

    // Obliquity (slant) factor.
    let f = 1.0 + 16.0 * (0.53 - e).powi(3);

    // Amplitude and period of the cosine model.
    let mut amp = coeffs.alpha[0]
        + coeffs.alpha[1] * phi_m
        + coeffs.alpha[2] * phi_m * phi_m
        + coeffs.alpha[3] * phi_m.powi(3);
    if amp < 0.0 {
        amp = 0.0;
    }
    let mut per = coeffs.beta[0]
        + coeffs.beta[1] * phi_m
        + coeffs.beta[2] * phi_m * phi_m
        + coeffs.beta[3] * phi_m.powi(3);
    if per < 72_000.0 {
        per = 72_000.0;
    }

    let x = 2.0 * PI * (t - 50_400.0) / per;
    // Delay in seconds (night term 5 ns plus the daytime cosine), then to metres.
    let t_iono = if x.abs() < 1.57 {
        f * (5e-9 + amp * (1.0 - x * x / 2.0 + x.powi(4) / 24.0))
    } else {
        f * 5e-9
    };
    t_iono * C_M_PER_S
}

// ----------------------------------------------------------------------------
// Troposphere — Saastamoinen zenith delay + Niell (1996) mapping function.
// ----------------------------------------------------------------------------

/// A simple standard-atmosphere meteorology at the receiver: pressure (hPa),
/// temperature (K), and relative humidity (fraction).
#[derive(Clone, Copy, Debug)]
pub struct Meteo {
    pub pressure_hpa: f64,
    pub temp_k: f64,
    pub humidity: f64,
}

impl Default for Meteo {
    /// US Standard Atmosphere at sea level, 50 % humidity.
    fn default() -> Self {
        Self {
            pressure_hpa: 1013.25,
            temp_k: 288.15,
            humidity: 0.5,
        }
    }
}

/// Saturation water-vapour partial pressure (hPa) at temperature `temp_k`
/// scaled by relative humidity — the Magnus/Tetens form.
fn water_vapour_pressure_hpa(meteo: &Meteo) -> f64 {
    let tc = meteo.temp_k - 273.15;
    let es = 6.108 * (17.15 * tc / (234.7 + tc)).exp();
    meteo.humidity * es
}

/// Saastamoinen zenith hydrostatic and wet delays (metres) at geodetic latitude
/// `lat_rad` and height `h_m`. (Davis et al. 1985; Groves §9.4.)
pub fn saastamoinen_zenith_m(meteo: &Meteo, lat_rad: f64, h_m: f64) -> (f64, f64) {
    let denom = 1.0 - 0.00266 * (2.0 * lat_rad).cos() - 0.00028 * (h_m / 1000.0);
    let zhd = 0.0022768 * meteo.pressure_hpa / denom;
    let e = water_vapour_pressure_hpa(meteo);
    let zwd = 0.0022768 * (1255.0 / meteo.temp_k + 0.05) * e / denom;
    (zhd, zwd)
}

/// Evaluate the Marini/Niell three-term continued-fraction mapping, normalised so
/// that `m(90°) = 1`.
fn marini_mapping(sin_e: f64, a: f64, b: f64, c: f64) -> f64 {
    let num = 1.0 + a / (1.0 + b / (1.0 + c));
    let den = sin_e + a / (sin_e + b / (sin_e + c));
    num / den
}

/// Linear interpolation of a Niell coefficient row by absolute latitude (deg),
/// with the table anchored at 15/30/45/60/75°.
fn niell_interp(table: &[f64; 5], lat_deg_abs: f64) -> f64 {
    let lats = [15.0, 30.0, 45.0, 60.0, 75.0];
    if lat_deg_abs <= lats[0] {
        return table[0];
    }
    if lat_deg_abs >= lats[4] {
        return table[4];
    }
    let i = lats.iter().position(|&l| lat_deg_abs < l).unwrap();
    let f = (lat_deg_abs - lats[i - 1]) / (lats[i] - lats[i - 1]);
    table[i - 1] + f * (table[i] - table[i - 1])
}

/// Niell (1996) hydrostatic mapping function at elevation `el_rad` for geodetic
/// latitude `lat_rad`, height `h_m`, and day-of-year `doy` (with a southern-
/// hemisphere half-year phase shift).
pub fn niell_hydrostatic(el_rad: f64, lat_rad: f64, h_m: f64, doy: f64) -> f64 {
    const AVG_A: [f64; 5] = [
        1.2769934e-3,
        1.2683230e-3,
        1.2465397e-3,
        1.2196049e-3,
        1.2045996e-3,
    ];
    const AVG_B: [f64; 5] = [
        2.9153695e-3,
        2.9152299e-3,
        2.9288445e-3,
        2.9022565e-3,
        2.9024912e-3,
    ];
    const AVG_C: [f64; 5] = [
        62.610505e-3,
        62.837393e-3,
        63.721774e-3,
        63.824265e-3,
        64.258455e-3,
    ];
    const AMP_A: [f64; 5] = [0.0, 1.2709626e-5, 2.6523662e-5, 3.4000452e-5, 4.1202191e-5];
    const AMP_B: [f64; 5] = [0.0, 2.1414979e-5, 3.0160779e-5, 7.2562722e-5, 11.723375e-5];
    const AMP_C: [f64; 5] = [0.0, 9.0128400e-5, 4.3497037e-5, 84.795348e-5, 170.37206e-5];
    const A_HT: f64 = 2.53e-5;
    const B_HT: f64 = 5.49e-3;
    const C_HT: f64 = 1.14e-3;

    let lat_abs = lat_rad.to_degrees().abs();
    // Seasonal phase: southern hemisphere is shifted half a year.
    let doy_adj = if lat_rad < 0.0 { doy + 182.625 } else { doy };
    let season = (2.0 * PI * (doy_adj - 28.0) / 365.25).cos();
    let coef = |avg: &[f64; 5], amp: &[f64; 5]| {
        niell_interp(avg, lat_abs) - niell_interp(amp, lat_abs) * season
    };
    let (a, b, c) = (
        coef(&AVG_A, &AMP_A),
        coef(&AVG_B, &AMP_B),
        coef(&AVG_C, &AMP_C),
    );
    let sin_e = el_rad.sin();
    let m = marini_mapping(sin_e, a, b, c);
    // Height correction (km): the difference between a 1/sin and the fixed-coeff
    // mapping, scaled by height.
    let ht = 1.0 / sin_e - marini_mapping(sin_e, A_HT, B_HT, C_HT);
    m + ht * (h_m / 1000.0)
}

/// Niell (1996) wet mapping function (latitude-only coefficients, no height term).
pub fn niell_wet(el_rad: f64, lat_rad: f64) -> f64 {
    const A: [f64; 5] = [
        5.8021897e-4,
        5.6794847e-4,
        5.8118019e-4,
        5.9727542e-4,
        6.1641693e-4,
    ];
    const B: [f64; 5] = [
        1.4275268e-3,
        1.5138625e-3,
        1.4572752e-3,
        1.5007428e-3,
        1.7599082e-3,
    ];
    const C: [f64; 5] = [
        4.3472961e-2,
        4.6729510e-2,
        4.3908931e-2,
        4.4626982e-2,
        5.4736038e-2,
    ];
    let lat_abs = lat_rad.to_degrees().abs();
    let (a, b, c) = (
        niell_interp(&A, lat_abs),
        niell_interp(&B, lat_abs),
        niell_interp(&C, lat_abs),
    );
    marini_mapping(el_rad.sin(), a, b, c)
}

/// Total slant tropospheric delay (metres): the Saastamoinen zenith hydrostatic
/// and wet delays, each projected with its Niell mapping function.
pub fn tropo_delay_m(meteo: &Meteo, lat_rad: f64, h_m: f64, el_rad: f64, doy: f64) -> f64 {
    let (zhd, zwd) = saastamoinen_zenith_m(meteo, lat_rad, h_m);
    zhd * niell_hydrostatic(el_rad, lat_rad, h_m, doy) + zwd * niell_wet(el_rad, lat_rad)
}

/// Euclidean distance between two ECEF points (m).
pub fn geometric_range_m(a: Vec3, b: Vec3) -> f64 {
    ((a[0] - b[0]).powi(2) + (a[1] - b[1]).powi(2) + (a[2] - b[2]).powi(2)).sqrt()
}

// ----------------------------------------------------------------------------
// Measurement-domain scenario pack.
// ----------------------------------------------------------------------------

fn default_mask_deg() -> f64 {
    5.0
}
fn default_p_fa() -> f64 {
    1e-5
}
fn default_p_md() -> f64 {
    1e-3
}
fn default_uere() -> f64 {
    3.0
}
fn default_alert_h() -> f64 {
    40.0
}
fn default_alert_v() -> f64 {
    50.0
}

/// The receiver: a geodetic position and a clock bias (expressed as a range
/// offset `c·δt_rx` in metres).
#[derive(Clone, Debug, Deserialize, Serialize)]
pub struct ReceiverCfg {
    pub lat_deg: f64,
    pub lon_deg: f64,
    #[serde(default)]
    pub alt_m: f64,
    /// Receiver clock bias as an equivalent range (m).
    #[serde(default)]
    pub clock_bias_m: f64,
}

/// The `[iono]` block: the eight Klobuchar coefficients and the GPS seconds-of-day.
#[derive(Clone, Debug, Deserialize, Serialize)]
pub struct IonoCfg {
    pub alpha: [f64; 4],
    pub beta: [f64; 4],
    /// GPS seconds of day at the start of the run (drives the diurnal cosine).
    #[serde(default = "default_gps_sod")]
    pub gps_seconds_of_day: f64,
}
fn default_gps_sod() -> f64 {
    50_400.0
}

/// The `[tropo]` block: surface meteorology and day-of-year.
#[derive(Clone, Debug, Deserialize, Serialize)]
pub struct TropoCfg {
    #[serde(default = "default_pressure")]
    pub pressure_hpa: f64,
    #[serde(default = "default_temp")]
    pub temp_k: f64,
    #[serde(default = "default_humidity")]
    pub humidity: f64,
    #[serde(default = "default_doy")]
    pub day_of_year: f64,
}
fn default_pressure() -> f64 {
    1013.25
}
fn default_temp() -> f64 {
    288.15
}
fn default_humidity() -> f64 {
    0.5
}
fn default_doy() -> f64 {
    180.0
}

/// A measurement-domain GNSS simulation: a Walker constellation seen by a ground
/// receiver, with modelled ionospheric and tropospheric delays, clock offsets,
/// thermal noise and multipath, fed to snapshot RAIM for protection levels.
#[derive(Clone, Debug, Deserialize, Serialize)]
pub struct GnssSimScenario {
    pub seed: u64,
    pub time: TimeCfg,
    pub receiver: ReceiverCfg,
    pub constellation: WalkerSgp4,
    #[serde(default)]
    pub iono: Option<IonoCfg>,
    #[serde(default)]
    pub tropo: Option<TropoCfg>,
    #[serde(default = "default_mask_deg")]
    pub mask_deg: f64,
    /// 1σ thermal-noise on each pseudorange (m). Set 0 for a noise-free run.
    #[serde(default)]
    pub noise_sigma_m: f64,
    /// Peak deterministic multipath error (m), scaled by `(1 − sin elevation)`.
    #[serde(default)]
    pub multipath_m: f64,
    /// Peak per-satellite clock offset (m); a deterministic per-PRN value.
    #[serde(default)]
    pub sat_clock_rms_m: f64,
    /// RAIM measurement error model (1σ user-equivalent range error, m).
    #[serde(default = "default_uere")]
    pub uere_m: f64,
    #[serde(default = "default_p_fa")]
    pub p_fa: f64,
    #[serde(default = "default_p_md")]
    pub p_md: f64,
    #[serde(default = "default_alert_h")]
    pub alert_limit_h_m: f64,
    #[serde(default = "default_alert_v")]
    pub alert_limit_v_m: f64,
}

/// One satellite's simulated measurement at one epoch.
#[derive(Clone, Debug, Serialize)]
pub struct SatMeasurement {
    pub prn: usize,
    pub el_deg: f64,
    pub az_deg: f64,
    pub pseudorange_m: f64,
    pub doppler_hz: f64,
    pub cn0_dbhz: f64,
    pub iono_correction_m: f64,
    pub tropo_correction_m: f64,
    pub sat_clock_m: f64,
}

/// The RAIM solution at one epoch.
#[derive(Clone, Debug, Serialize)]
pub struct RaimEpoch {
    pub n_used: usize,
    pub fault_detected: bool,
    pub hpl_m: f64,
    pub vpl_m: f64,
    pub test_statistic: f64,
    pub threshold: f64,
}

/// One epoch of the measurement simulation.
#[derive(Clone, Debug, Serialize)]
pub struct GnssSimEpoch {
    pub t: f64,
    pub n_visible: usize,
    pub raim: Option<RaimEpoch>,
    pub measurements: Vec<SatMeasurement>,
}

/// Figures of merit for the measurement simulation.
#[derive(Clone, Debug, Serialize)]
pub struct GnssSimFoM {
    /// Fraction of epochs where RAIM is available (≥5 satellites) and both
    /// protection levels are under their alert limits.
    pub raim_availability: f64,
    pub mean_hpl_m: f64,
    pub mean_vpl_m: f64,
    /// Fraction of epochs where RAIM declared a fault.
    pub fault_rate: f64,
    pub mean_iono_m: f64,
    pub mean_tropo_m: f64,
}

/// The measurement-domain GNSS simulation result.
#[derive(Clone, Debug, Serialize)]
pub struct GnssSimResult {
    pub schema_version: String,
    pub engine_version: String,
    pub scenario_hash: String,
    pub seed: u64,
    pub fom: GnssSimFoM,
    pub gnss_measurements: Vec<GnssSimEpoch>,
}

/// A representative elevation-dependent C/N₀ (dB-Hz): ~48 at zenith, ~40 at the
/// horizon.
fn cn0_dbhz(el_rad: f64) -> f64 {
    40.0 + 8.0 * el_rad.sin()
}

/// A deterministic per-PRN satellite clock offset (m) in `[−rms, rms]`.
fn sat_clock_m(prn: usize, rms: f64) -> f64 {
    // A fixed irrational stride spreads the PRNs across the interval.
    let frac = (prn as f64 * 0.61803398875).fract();
    rms * (2.0 * frac - 1.0)
}

fn hash_scenario(scn: &GnssSimScenario) -> String {
    let c = serde_json::to_string(scn).expect("scenario serializes");
    let mut h = Sha256::new();
    h.update(c.as_bytes());
    hex::encode(h.finalize())
}

/// Run the measurement-domain GNSS simulation.
pub fn run_gnss_sim(scn: &GnssSimScenario) -> GnssSimResult {
    let station = Geodetic {
        lat_rad: scn.receiver.lat_deg.to_radians(),
        lon_rad: scn.receiver.lon_deg.to_radians(),
        alt_m: scn.receiver.alt_m,
    };
    let station_ecef = geodetic_to_ecef(station);
    let sats = scn.constellation.satellites();
    let lambda_l1 = C_M_PER_S / L1_HZ;
    let iono = scn.iono.as_ref().map(|c| {
        (
            KlobucharCoeffs {
                alpha: c.alpha,
                beta: c.beta,
            },
            c.gps_seconds_of_day,
        )
    });
    let (meteo, doy) = match &scn.tropo {
        Some(t) => (
            Some(Meteo {
                pressure_hpa: t.pressure_hpa,
                temp_k: t.temp_k,
                humidity: t.humidity,
            }),
            t.day_of_year,
        ),
        None => (None, 0.0),
    };

    let dt = scn.time.step_s;
    let n = (scn.time.duration_s / dt).round() as usize;
    let mut rng = ChaCha8Rng::seed_from_u64(scn.seed);
    let noise = Normal::new(0.0, scn.noise_sigma_m.max(0.0).max(1e-12)).unwrap();

    let sat_ecef_at = |i: usize, t: f64| -> Vec3 {
        teme_to_ecef(sats[i].position_eci(t), walker_epoch_jd() + t / 86_400.0)
    };

    let mut epochs = Vec::with_capacity(n + 1);
    let (mut avail, mut faults, mut hpl_sum, mut vpl_sum, mut hvpl_count) =
        (0usize, 0usize, 0.0, 0.0, 0usize);
    let (mut iono_sum, mut tropo_sum, mut meas_count) = (0.0, 0.0, 0usize);

    for k in 0..=n {
        let t = k as f64 * dt;
        let mut measurements = Vec::new();
        let mut raim_sats: Vec<Vec3> = Vec::new();
        let mut raim_resid: Vec<f64> = Vec::new();
        for (prn, _) in sats.iter().enumerate() {
            let sat = sat_ecef_at(prn, t);
            if !is_visible(station, sat, scn.mask_deg) {
                continue;
            }
            let look = look_angles(station, sat);
            let (el, az) = (look.el_rad, look.az_rad);
            let r_geom = geometric_range_m(station_ecef, sat);
            let i_delay = iono
                .as_ref()
                .map(|(c, sod)| {
                    klobuchar_delay_m(c, station.lat_rad, station.lon_rad, el, az, sod + t)
                })
                .unwrap_or(0.0);
            let t_delay = meteo
                .as_ref()
                .map(|m| tropo_delay_m(m, station.lat_rad, station.alt_m, el, doy))
                .unwrap_or(0.0);
            let sclk = sat_clock_m(prn, scn.sat_clock_rms_m);
            let multipath = scn.multipath_m * (1.0 - el.sin());
            let eps = if scn.noise_sigma_m > 0.0 {
                noise.sample(&mut rng)
            } else {
                0.0
            };
            // The modelled pseudorange.
            let pr =
                r_geom + scn.receiver.clock_bias_m - sclk + i_delay + t_delay + multipath + eps;

            // Doppler: central-difference the geometric range, convert to Hz.
            let d = 0.5_f64.min(dt.max(1e-3) * 0.5);
            let r_plus = geometric_range_m(station_ecef, sat_ecef_at(prn, t + d));
            let r_minus = geometric_range_m(station_ecef, sat_ecef_at(prn, (t - d).max(0.0)));
            let range_rate = (r_plus - r_minus) / (2.0 * d);
            let doppler_hz = -range_rate / lambda_l1;

            // RAIM residual: observed pseudorange minus the predicted range built
            // from the known geometry, the known receiver/satellite clocks, and the
            // applied iono/tropo corrections — i.e. the unmodelled error (noise +
            // multipath). A perfectly-modelled measurement has a zero residual.
            let predicted = r_geom + scn.receiver.clock_bias_m - sclk + i_delay + t_delay;
            raim_sats.push(sat);
            raim_resid.push(pr - predicted);

            iono_sum += i_delay;
            tropo_sum += t_delay;
            meas_count += 1;
            measurements.push(SatMeasurement {
                prn,
                el_deg: el.to_degrees(),
                az_deg: az.to_degrees(),
                pseudorange_m: pr,
                doppler_hz,
                cn0_dbhz: cn0_dbhz(el),
                iono_correction_m: i_delay,
                tropo_correction_m: t_delay,
                sat_clock_m: sclk,
            });
        }

        let n_visible = measurements.len();
        let raim = crate::raim::snapshot_raim(
            station_ecef,
            &raim_sats,
            &raim_resid,
            scn.uere_m,
            scn.p_fa,
            scn.p_md,
        )
        .map(|r| RaimEpoch {
            n_used: r.n_used,
            fault_detected: r.fault_detected,
            hpl_m: r.hpl_m,
            vpl_m: r.vpl_m,
            test_statistic: r.test_statistic,
            threshold: r.threshold,
        });
        if let Some(r) = &raim {
            hpl_sum += r.hpl_m;
            vpl_sum += r.vpl_m;
            hvpl_count += 1;
            if r.fault_detected {
                faults += 1;
            }
            if r.hpl_m <= scn.alert_limit_h_m && r.vpl_m <= scn.alert_limit_v_m {
                avail += 1;
            }
        }
        epochs.push(GnssSimEpoch {
            t,
            n_visible,
            raim,
            measurements,
        });
    }

    let denom = (n + 1) as f64;
    let hv = hvpl_count.max(1) as f64;
    let mc = meas_count.max(1) as f64;
    GnssSimResult {
        schema_version: crate::interchange::SCHEMA_VERSION.into(),
        engine_version: env!("CARGO_PKG_VERSION").into(),
        scenario_hash: hash_scenario(scn),
        seed: scn.seed,
        fom: GnssSimFoM {
            raim_availability: avail as f64 / denom,
            mean_hpl_m: hpl_sum / hv,
            mean_vpl_m: vpl_sum / hv,
            fault_rate: faults as f64 / denom,
            mean_iono_m: iono_sum / mc,
            mean_tropo_m: tropo_sum / mc,
        },
        gnss_measurements: epochs,
    }
}

/// Render the horizontal and vertical protection levels over time against their
/// alert limits.
pub fn to_svg(result: &GnssSimResult, alert_h_m: f64, alert_v_m: f64) -> String {
    let (w, h) = (820.0_f64, 420.0_f64);
    let (ml, mr, mt, mb) = (60.0_f64, 20.0_f64, 30.0_f64, 50.0_f64);
    let pw = w - ml - mr;
    let ph = h - mt - mb;
    let ep = &result.gnss_measurements;
    let t_max = ep.iter().map(|e| e.t).fold(1.0_f64, f64::max);
    let mut y_max = alert_v_m.max(alert_h_m);
    for e in ep {
        if let Some(r) = &e.raim {
            y_max = y_max.max(r.hpl_m).max(r.vpl_m);
        }
    }
    y_max *= 1.15;
    if y_max <= 0.0 {
        y_max = 1.0;
    }
    let xof = |t: f64| ml + (t / t_max) * pw;
    let yof = |v: f64| mt + ph - (v.min(y_max) / y_max) * ph;
    let poly = |sel: &dyn Fn(&RaimEpoch) -> f64| {
        ep.iter()
            .filter_map(|e| {
                e.raim
                    .as_ref()
                    .map(|r| format!("{:.1},{:.1}", xof(e.t), yof(sel(r))))
            })
            .collect::<Vec<_>>()
            .join(" ")
    };
    let axis_y = mt + ph;
    let mut svg = String::new();
    svg.push_str(&format!("<svg xmlns=\"http://www.w3.org/2000/svg\" width=\"{w:.0}\" height=\"{h:.0}\" font-family=\"sans-serif\" font-size=\"12\" fill=\"#bcb3a3\"><rect width=\"{w:.0}\" height=\"{h:.0}\" fill=\"#0c0b08\"/>"));
    svg.push_str(&format!("<text x=\"{ml:.0}\" y=\"18\" font-size=\"15\" font-weight=\"bold\">RAIM protection levels vs alert limits</text>"));
    svg.push_str(&crate::chart::y_axis(
        ml,
        mt,
        pw,
        ph,
        y_max,
        "protection level (m)",
    ));
    svg.push_str(&format!("<line x1=\"{ml:.0}\" y1=\"{mt:.0}\" x2=\"{ml:.0}\" y2=\"{axis_y:.0}\" stroke=\"#342c21\"/><line x1=\"{ml:.0}\" y1=\"{axis_y:.0}\" x2=\"{:.0}\" y2=\"{axis_y:.0}\" stroke=\"#342c21\"/>", ml + pw));
    let right = ml + pw;
    let val_y = yof(alert_v_m);
    let val_label_x = ml + 4.0;
    let val_label_y = val_y - 4.0;
    svg.push_str(&format!("<line x1=\"{ml:.0}\" y1=\"{val_y:.1}\" x2=\"{right:.0}\" y2=\"{val_y:.1}\" stroke=\"#e5645a\" stroke-dasharray=\"6 4\"/><text x=\"{val_label_x:.0}\" y=\"{val_label_y:.1}\" fill=\"#e5645a\">VAL {alert_v_m:.0} m</text>"));
    svg.push_str(&format!(
        "<polyline fill=\"none\" stroke=\"#e0bd84\" stroke-width=\"2\" points=\"{}\"/>",
        poly(&|r| r.hpl_m)
    ));
    svg.push_str(&format!(
        "<polyline fill=\"none\" stroke=\"#c79e63\" stroke-width=\"2\" points=\"{}\"/>",
        poly(&|r| r.vpl_m)
    ));
    svg.push_str(&format!("<text x=\"{:.0}\" y=\"44\" fill=\"#e0bd84\">HPL</text><text x=\"{:.0}\" y=\"60\" fill=\"#c79e63\">VPL</text>", ml + 10.0, ml + 10.0));
    svg.push_str("</svg>");
    svg
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn klobuchar_is_positive_and_largest_at_low_elevation() {
        let c = KlobucharCoeffs::default();
        let (lat, lon) = (40f64.to_radians(), -100f64.to_radians());
        // Mid-afternoon local-ish; a high-elevation satellite.
        let high = klobuchar_delay_m(&c, lat, lon, 80f64.to_radians(), 0.0, 50_400.0);
        let low = klobuchar_delay_m(&c, lat, lon, 10f64.to_radians(), 0.0, 50_400.0);
        assert!(high > 0.0, "delay should be positive, got {high}");
        // The obliquity factor (≈2.7× from 80°→10°) makes a low satellite's slant
        // delay substantially larger, though the amplitude term varies with the
        // pierce-point geomagnetic latitude so the net ratio is < the raw obliquity.
        assert!(
            low > high * 1.5,
            "low-el {low} should exceed high-el {high} by ≥1.5×"
        );
        // Daytime L1 iono is a few metres to tens of metres at the zenith.
        assert!(
            high > 0.5 && high < 30.0,
            "zenith-ish delay {high} m out of range"
        );
    }

    #[test]
    fn klobuchar_night_floor_is_five_ns() {
        // Far from the 14:00 LT peak and at high elevation (F≈1), the model floors
        // at the 5 ns night term: ~1.5 m. Use a longitude/time putting LT near 02:00.
        let c = KlobucharCoeffs::default();
        let d = klobuchar_delay_m(&c, 0.0, 0.0, 90f64.to_radians(), 0.0, 7_200.0);
        let night_m = 5e-9 * C_M_PER_S; // ≈ 1.499 m
        assert!(
            (d - night_m).abs() < 0.2,
            "night delay {d} should be ≈ {night_m}"
        );
    }

    #[test]
    fn klobuchar_agrees_with_rtklib_reference_implementation() {
        // Cross-implementation check of the full IS-GPS-200 §20.3.3.5.2.5 broadcast
        // model against RTKLIB's `ionmodel()` (`src/rtkcmn.c`), the de-facto
        // open-source GNSS reference implementation of the *same* semicircle
        // algorithm. There is no official published numeric test vector for the
        // broadcast Klobuchar model (IS-GPS-200, the GPS SPS spec, and the ESA GNSS
        // book all give the algorithm but no worked example — and the ESA book uses
        // a different, rigorous spherical-trig form), so this pins agreement with an
        // independent reference implementation, not against measured truth.
        //
        // Inputs: user (40°N, 260°E), satellite at E=20°, A=210°, GPS time of week
        // 518400 s; coefficients are RTKLIB's built-in default broadcast set
        // (`rtkcmn.c`, dated 2004/1/1). RTKLIB's `ionmodel()` returns 6.1278 m of L1
        // slant delay for this case; Kshana reproduces it to sub-millimetre.
        let coeffs = KlobucharCoeffs {
            alpha: [0.1118e-7, -0.7451e-8, -0.5961e-7, 0.1192e-6],
            beta: [0.1167e6, -0.2294e6, -0.1311e6, 0.1049e7],
        };
        let delay = klobuchar_delay_m(
            &coeffs,
            40f64.to_radians(),
            260f64.to_radians(),
            20f64.to_radians(),
            210f64.to_radians(),
            518_400.0,
        );
        const RTKLIB_REF_M: f64 = 6.1278;
        assert!(
            (delay - RTKLIB_REF_M).abs() < 1e-3,
            "Klobuchar L1 delay {delay} m should match RTKLIB ionmodel() {RTKLIB_REF_M} m"
        );
    }

    #[test]
    fn saastamoinen_zenith_is_about_2_3_m_at_sea_level() {
        let m = Meteo::default();
        let (zhd, zwd) = saastamoinen_zenith_m(&m, 45f64.to_radians(), 0.0);
        // Hydrostatic ≈ 2.3 m, wet ≈ 0.1–0.2 m; total ≈ 2.4–2.5 m.
        assert!((zhd - 2.30).abs() < 0.05, "ZHD {zhd}");
        assert!(zwd > 0.05 && zwd < 0.30, "ZWD {zwd}");
    }

    #[test]
    fn niell_mapping_is_unity_at_zenith_and_grows_toward_the_horizon() {
        let lat = 45f64.to_radians();
        assert!((niell_hydrostatic(PI / 2.0, lat, 0.0, 100.0) - 1.0).abs() < 1e-9);
        assert!((niell_wet(PI / 2.0, lat) - 1.0).abs() < 1e-9);
        // Monotone increase as elevation drops; ~10 at 5.7° for hydrostatic.
        let m30 = niell_hydrostatic(30f64.to_radians(), lat, 0.0, 100.0);
        let m5 = niell_hydrostatic(5f64.to_radians(), lat, 0.0, 100.0);
        assert!(m30 > 1.9 && m30 < 2.1, "m(30°)={m30}");
        assert!(m5 > 9.0 && m5 < 12.0, "m(5°)={m5}");
        assert!(m5 > m30);
    }

    #[test]
    fn total_tropo_at_30_deg_is_about_twice_the_zenith() {
        let meteo = Meteo::default();
        let lat = 45f64.to_radians();
        let (zhd, zwd) = saastamoinen_zenith_m(&meteo, lat, 0.0);
        let zenith_total = zhd + zwd;
        let slant30 = tropo_delay_m(&meteo, lat, 0.0, 30f64.to_radians(), 100.0);
        assert!(
            (slant30 / zenith_total - 2.0).abs() < 0.05,
            "slant/zenith = {}",
            slant30 / zenith_total
        );
    }

    // --- measurement-domain pack ---

    fn gps_like() -> WalkerSgp4 {
        WalkerSgp4 {
            altitude_km: 20_200.0,
            inclination_deg: 55.0,
            planes: 6,
            sats_per_plane: 4,
            phasing_f: 1.0,
        }
    }

    fn sim_scenario(noise_sigma_m: f64) -> GnssSimScenario {
        GnssSimScenario {
            seed: 42,
            time: TimeCfg {
                step_s: 30.0,
                duration_s: 300.0,
            },
            receiver: ReceiverCfg {
                lat_deg: 40.0,
                lon_deg: -3.0,
                alt_m: 600.0,
                clock_bias_m: 1234.5,
            },
            constellation: gps_like(),
            iono: Some(IonoCfg {
                alpha: [3.82e-8, 1.49e-8, -1.79e-7, 0.0],
                beta: [1.43e5, 0.0, -3.28e5, 1.13e5],
                gps_seconds_of_day: 50_400.0,
            }),
            tropo: Some(TropoCfg {
                pressure_hpa: 1013.25,
                temp_k: 288.15,
                humidity: 0.5,
                day_of_year: 180.0,
            }),
            mask_deg: 5.0,
            noise_sigma_m,
            multipath_m: 0.0,
            sat_clock_rms_m: 30.0,
            uere_m: 3.0,
            p_fa: 1e-5,
            p_md: 1e-3,
            alert_limit_h_m: 40.0,
            alert_limit_v_m: 50.0,
        }
    }

    #[test]
    fn zero_noise_pseudorange_matches_geometry_plus_corrections_to_a_millimetre() {
        // The milestone's acceptance: with no noise or multipath, the simulated
        // pseudorange equals geometric range + receiver clock − sat clock + iono +
        // tropo exactly. The RAIM residual (observed − fully-modelled prediction)
        // is therefore zero to sub-millimetre at every satellite of every epoch.
        let r = run_gnss_sim(&sim_scenario(0.0));
        let mut checked = 0;
        for e in &r.gnss_measurements {
            for m in &e.measurements {
                // Reconstruct the prediction from the reported correction fields.
                // pseudorange − (clock + iono + tropo − sat_clock) must be the bare
                // geometric range, which we cannot recompute here without the sat
                // position, so instead assert the RAIM residual is ~0 (below).
                assert!(
                    m.pseudorange_m > 1.9e7,
                    "pseudorange {} implausible",
                    m.pseudorange_m
                );
                checked += 1;
            }
            // Every modelled term is accounted for ⇒ residual ≈ 0 ⇒ test stat ≈ 0.
            if let Some(raim) = &e.raim {
                assert!(
                    raim.test_statistic < 1e-6,
                    "zero-noise RAIM statistic should be ~0, got {}",
                    raim.test_statistic
                );
                assert!(!raim.fault_detected);
            }
        }
        assert!(checked > 20, "expected many measurements, got {checked}");
    }

    #[test]
    fn corrections_are_reported_and_physically_sized() {
        let r = run_gnss_sim(&sim_scenario(0.0));
        let mut saw = false;
        for e in &r.gnss_measurements {
            for m in &e.measurements {
                // Iono: ~1–30 m slant; tropo: ~2–15 m slant (zenith ~2.4 m × mapping).
                assert!(m.iono_correction_m > 0.0 && m.iono_correction_m < 60.0);
                assert!(m.tropo_correction_m > 2.0 && m.tropo_correction_m < 30.0);
                assert!(m.cn0_dbhz > 39.0 && m.cn0_dbhz < 49.0);
                // GPS line-of-sight Doppler is within roughly ±5 kHz on L1.
                assert!(m.doppler_hz.abs() < 6000.0, "doppler {}", m.doppler_hz);
                saw = true;
            }
        }
        assert!(saw);
        assert!(r.fom.mean_iono_m > 0.0 && r.fom.mean_tropo_m > 2.0);
    }

    #[test]
    fn raim_produces_protection_levels_and_run_is_deterministic() {
        let r = run_gnss_sim(&sim_scenario(1.0));
        let with_raim = r
            .gnss_measurements
            .iter()
            .filter(|e| e.raim.is_some())
            .count();
        assert!(
            with_raim > 0,
            "RAIM should be available with a full constellation"
        );
        for e in &r.gnss_measurements {
            if let Some(raim) = &e.raim {
                assert!(raim.hpl_m > 0.0 && raim.vpl_m > 0.0);
                assert!(raim.n_used >= 5);
            }
        }
        // Deterministic.
        let a = serde_json::to_string(&run_gnss_sim(&sim_scenario(1.0))).unwrap();
        let b = serde_json::to_string(&run_gnss_sim(&sim_scenario(1.0))).unwrap();
        assert_eq!(a, b);
    }

    #[test]
    fn disabling_the_atmosphere_zeroes_the_corrections() {
        let mut scn = sim_scenario(0.0);
        scn.iono = None;
        scn.tropo = None;
        let r = run_gnss_sim(&scn);
        for e in &r.gnss_measurements {
            for m in &e.measurements {
                assert_eq!(m.iono_correction_m, 0.0);
                assert_eq!(m.tropo_correction_m, 0.0);
            }
        }
    }
}