Jamming, spoofing and the spectrum (scenario kind family)

6 scenario kinds of the engine v0.38.0. Advisory: analysis output from a research and engineering tool, not type-approved navigation equipment and not a safety-of-life service. Every label here is the engine's or the verification matrix's; a kind is MODELLED unless the matrix says otherwise (how a label is earned: the README's Evidence section).

Topic pages: SPECTRUM, DECEIVED-TIME, IQ-FOLLOW. Run one: kshana scenarios/jamming-demo.toml. All families: KINDS.md.

jamming#

Link-budget jamming: J/S → effective C/N₀ → loss of lock.

Required fields: seed, time, receiver, constellation. Optional: jammer, mask_deg, tracking_threshold_dbhz, degraded_margin_db, signal_power_dbw, temp_k, freq_hz, chip_rate_hz, skyline.

Ledger: MODELLED. Full page, with the inputs, results, evidence and limitations: jamming.

kshana scenarios/jamming-demo.toml

spoof#

Stochastic time-spoof detector (Neyman–Pearson / χ²₁) with Monte-Carlo P_fa/P_md.

Required fields: threshold_ns, time, attack, clock_quantum, clock_classical.

Ledger: MODELLED. Full page, with the inputs, results, evidence and limitations: spoof.

kshana scenarios/spoof-attack.toml

spoof-detect#

Combined RF/measurement spoof detector (multi-SV RAIM-consistency + AGC + SQM, fused) vs a parameterised attack (power advantage, carrier-phase alignment, time/position push; TEXBAT-style).

Required fields: attack. Optional: satellites, detector.

Ledger: MODELLED. Full page, with the inputs, results, evidence and limitations: spoof-detect.

kshana scenarios/spoof-detect.toml

tracking-loop#

Tracking-loop loss of lock: C/N0 -> lock/unlock WITH HYSTERESIS, the time it takes to lose lock, and spoof pull-in as a function of code and carrier offset RATE. The engine's existing denial and capture criteria are ratios on received power (a jammer denies at J/S = 30 dB, a spoofer captures at 3 dB); real loss of lock is not a power ratio, it is where a loop's jitter plus dynamic stress leaves its tracking-threshold budget. Reports (1) carrier (Costas) and code (non-coherent early/late) 1-sigma thermal jitter versus C/N0, squaring loss included, against the stated rules 3sigma_PLL + theta_e <= 45 deg (the 15-degree rule in three-sigma form) and 3sigma_DLL + ramp lag <= d/2 chips; (2) the drop and re-lock C/N0 thresholds with the binding loop named, the re-lock threshold DERIVED from the wider bandwidth a receiver re-pulls-in at rather than asserted, so the hysteresis width must fall between 5log10(ratio) and 10log10(ratio) dB; (3) the declared time to lose lock from a two-threshold lock detector with confirmation dwells driven over a C/N0 ramp, reported separately from the physical phase-escape time (the Viterbi mean time between cycle slips, in log10 s because it spans hundreds of decades); (4) spoof pull-in limits — the largest code slew and carrier Doppler rate a victim's loops can follow, with a map over both axes, because a spoofer that slews faster than these does not capture the loop, it outruns it; and (5) the DENIAL RADIUS the loop dynamics imply reported ALONGSIDE the existing power-ratio radius, never in place of it, with their signed difference as its own named field (denial_radius_delta_km) and a definition beside it. Every emitted numeric field carries a unit and a provenance class. VALIDATED closed forms: the carrier and code jitter expressions, the loop-filter reduction sqrt(2B_nT) and the first-order ramp lag are cross-checked against the engine's own sdr correlator stepped forward on seeded synthetic IF — a different route to the same numbers — to 0.6 %, 3.7 % and 0.17 % respectively, and the modified Bessel I0 behind the cycle-slip time against standard tabulated values to 1e-7. The lower validity limit is asserted, not merely stated: below ~32 dB-Hz at 1 ms the real atan discriminator saturates against its +-pi/2 range and measures less jitter than any linear theory. MODELLED: the loop bandwidths, integration time, correlator spacing, pull-in bandwidth ratio, confirmation dwells, jammer power and antenna gains are representative band figures, not a datasheet. NOT modelled: oscillator (Allan-deviation) and vibration jitter, front-end bandwidth limiting, multipath, AGC dynamics, data-bit-transition loss, external aiding, and any half-cycle correction to the Costas cycle-slip formula. Under loop dynamics spoof capture has no radius at all — the binding constraint is offset rate, not power — so that field is null with a reason rather than a fabricated number. Not a certified receiver-performance product.

Required fields: none. Optional: pll_bandwidth_hz, dll_bandwidth_hz, predetection_integration_s, correlator_spacing_chips, carrier_allowance_deg, code_allowance_chips, pullin_bandwidth_ratio, drop_confirm_epochs, relock_confirm_epochs, doppler_rate_hz_per_s, code_slew_chips_per_s, cn0_high_dbhz, cn0_low_dbhz, ramp_duration_s, jitter_table_cn0_dbhz, jammer_power_dbw, jammer_gain_dbi, rx_gain_toward_jammer_db, rx_gain_toward_sat_db, signal_power_dbw, temp_k, freq_hz, chip_rate_hz, jammer_type, q_override, denial_js_threshold_db, capture_js_threshold_db, spoof_code_slew_chips_per_s, spoof_doppler_rate_hz_per_s.

Ledger: MODELLED. Full page, with the inputs, results, evidence and limitations: tracking-loop.

kshana scenarios/tracking-loop.toml

spectrum#

L-band spectrum model and waterfall: the whole GNSS L band as one power spectral density (PSD), frequency across, time down, over a scripted jammer timeline. SIGNALS: GPS L1 C/A and L2C (BPSK(1)), Galileo E1 open service (MBOC(6,1,1/11), or BOC(1,1)), GPS L5 and Galileo E5a (BPSK(10)), each at its carrier with the interface-specification minimum received power (IS-GPS-200, IS-GPS-705, Galileo OS SIS ICD) unless overridden; GLONASS, BeiDou and Galileo E6 are not modelled. JAMMERS: continuous-wave tone, flat narrowband noise, linear sawtooth chirp and broadband noise matched to a band's own modulation, each with a centre, a bandwidth, a received power (direct, or EIRP and range through the jamming kind's free-space path loss) and on/off times. Each jammer is scored against each band by its spectral separation coefficient kappa over the receiver bandwidth, and (C/N0)_eff = [1/(C/N0) + sum (J/S) kappa]^-1; for one jammer this equals the jamming kind's anti-jam equation with Q = 1/(R_c kappa), and the report carries that cross-check (and the jamming kind's representative-Q figure beside it). The noise floor is k T_sys with T_sys = T_ant + 290 K (F - 1). The waterfall averages each cell over its bin and its row (chirps exactly, whole sweeps plus the partial one; jammers switching mid-row by their duty), so the C/N0 timeline and the picture come from the same numbers; result.json carries the grid block-averaged in power, per-band C/N0 and J/S per row, and per jammer per band the SSC, Q, J/S, in-band J/S and steady C/N0. Optional [iq]: the model drawn as complex IQ at one instant, written to and read back from a SigMF recording (cf32_le or ci16_le), and compared with a Welch estimate. Optional [recording] (native builds): a real SigMF recording's Welch PSD beside the model. VALIDATED: the signal PSDs against the textbook main-lobe widths (BPSK(n) 2n x 1.023 MHz null to null; BOC(1,1) lobes centred at +/-1.023 MHz) and the spectral separation coefficients against their Parseval closed forms and Kaplan & Hegarty's Q = 1 (CW) and 1.5 (matched). MODELLED: jammer powers, timeline, front-end bandwidths, the continuous-spectrum (no code lines) treatment and the absence of AGC, blanking and antenna pattern.

Required fields: none. Optional: seed, duration_s, step_s, receiver, grid, bands, jammers, iq, recording.

Ledger: MODELLED 3 · VALIDATED 2. Full page, with the inputs, results, evidence and limitations: spectrum.

kshana scenarios/l-band-waterfall-jamming.toml

impairment-eval#

Machine-learning RF-impairment detection evaluation testbed (13494): generate a labelled, parameter-grounded SYNTHETIC corpus (nominal/jamming/spoof-time/spoof-position/multipath), score a detector (energy|agc|sqm|parity|fused) with the detector-agnostic harness, and report AUC/ROC/confusion + per-class Pd at a target Pfa, plus the in- vs out-of-distribution optimism gap. MODELLED operating characteristics only — never field/IQ, no good/bad verdict.

Required fields: none. Optional: seed, n_per_class, nominal_cn0_dbhz, meas_noise, detector, target_pfa, shift_severity_scale, optimism_tol.

Ledger: MODELLED 1 · VALIDATED 2. Full page, with the inputs, results, evidence and limitations: impairment-eval.

kshana scenarios/impairment-eval.toml