Parameter set v1-default (algorithm version 1)
Selected in: phase-03 (normalisation), phase-04 (GPS validation calibration), phase-06 (speed filters); IMU filter values remain provisional until Phase 8.
Justification: docs/proofs/phase-04/qc_report.md, docs/proofs/phase-06/filter_selection.md.
Changing any value requires a new param_set_id and an algorithm_version bump; every artifact records the id and the full dictionary in its provenance.
Normalisation (Phase 3)
| parameter | value | meaning |
|---|---|---|
earth_radius_m |
6.37101e+06 | Earth radius for the local east-north projection |
stationary_speed_max_ms |
0.5 | GPS speed below which a second counts as stationary (up-vector calibration) |
stationary_min_seconds |
30 | stationary data needed for a confident up-vector |
stationary_gyro_max_rads |
0.05 | IMU-only stationary detection: gyro magnitude limit |
stationary_accel_std_max_ms2 |
0.3 | IMU-only stationary detection: accelerometer std limit |
timing_drift_fallback_s |
0.5 | cumulative-index timing falls back to per-payload timing above this residual |
chapter_gap_min_s |
0.5 | minimum STMP gap between consecutive chapters |
chapter_gap_max_s |
1.5 | maximum STMP gap between consecutive chapters |
GPS validation (Phase 4)
| parameter | value | meaning |
|---|---|---|
gps_fix_min |
3 | minimum fix quality (3 = 3-D fix) |
gps_dop_max |
3 | DOP above which a sample is invalid |
gps_dop_confidence_taper |
2 | DOP above which confidence is halved |
gps_prelock_valid_run_s |
2 | length of the first valid run that ends the pre-lock period |
gps_gap_s |
1 | gap length that splits the data into runs |
gps_gap_factor |
3 | a gap is dt > factor / nominal rate |
gps_stationary_speed_ms |
0.5 | speed below which duplicate positions are a receiver hold |
gps_moving_speed_ms |
3 | speed above which a duplicate position is a jump |
gps_jump_speed_factor |
2 | position-jump limit: factor × speed × dt |
gps_jump_noise_m |
3 | position-jump limit: added noise allowance |
gps_jump_speed_floor_ms |
60 | position-jump limit floor speed |
gps_speed_max_ms |
80 | reported speed above which a sample is a spike |
gps_spike_accel_ms2 |
15 | acceleration of a speed step that counts as a spike |
gps_spike_recovery_samples |
3 | samples within which a spike must reverse |
gps_hampel_window |
7 | Hampel window on reported speed |
gps_hampel_nsigma |
3 | Hampel threshold in robust sigmas |
gps_hampel_min_ms |
2 | Hampel minimum absolute deviation |
gps_implausible_accel_ms2 |
12 | sustained acceleration that lowers confidence |
gps_implausible_accel_s |
0.5 | duration for the sustained-acceleration rule |
gps_valid_fraction_min |
0.8 | session blocked below this valid fraction |
gps_min_valid_run_s |
60 | session blocked without a valid run this long |
gps_speed_consistency_window_s |
2 | window for the position-vs-reported speed witness |
gps_speed_consistency_max_ms |
1.5 | allowed disagreement of the speed witness |
gps_speed_consistency_fraction |
0.95 | fraction of moving windows that must agree |
Lap detection (Phase 5)
| parameter | value | meaning |
|---|---|---|
lap_heading_smooth_samples |
5 | moving-average window for the heading from positions |
lap_heading_min_speed_ms |
2 | heading undefined below this speed |
lap_race_min_speed_ms |
8 | samples faster than this define the racing surface |
lap_cross_min_speed_ms |
3 | minimum speed for a gate crossing |
lap_gate_min_halfwidth_m |
15 | minimum gate half-width |
lap_gate_spread_factor |
3 | gate half-width = factor × median lateral spread |
lap_gate_heading_max_deg |
45 | maximum heading error for a crossing |
gate_point_max_offset_m |
30 | a start/finish line moved by the racer must lie within this distance of the centerline |
lap_min_lap_s |
6 | minimum time between accepted crossings |
lap_progress_min_fraction |
0.95 | a lap must cover this fraction of the loop |
lap_backward_jump_m |
10 | maximum backward step in progress |
lap_distance_tolerance |
0.1 | integrated distance tolerance vs loop length |
lap_gps_valid_fraction_min |
0.9 | lap invalid below this valid fraction |
lap_max_gap_s |
1 | lap invalid with a GPS gap longer than this |
lap_pit_speed_ms |
2 | speed below which the car counts as stopped |
lap_pit_min_s |
2 | stopped time that makes a lap a pit lap |
lap_caution_factor |
1.35 | lap time factor over the fastest lap for caution |
lap_caution_speed_fraction |
0.7 | mean-speed fraction under the best for caution |
lap_implausibly_fast_factor |
0.95 | lap time must exceed factor × L / v_max |
Speed and path acceleration (Phase 6)
| parameter | value | meaning |
|---|---|---|
speed_cutoff_hz |
1 | zero-phase Butterworth cutoff on GPS speed |
speed_gap_fill_s |
0.5 | gaps up to this length are linearly filled |
sg_window_long_s |
0.7 | Savitzky–Golay window for longitudinal acceleration |
sg_window_heading_s |
1 | Savitzky–Golay window for the path heading |
sg_polyorder |
2 | Savitzky–Golay polynomial order |
brake_threshold_ms2 |
2 | deceleration that defines the brake point |
brake_sustain_s |
0.3 | how long braking must be sustained |
accel_threshold_ms2 |
1 | acceleration that defines the acceleration point |
accel_sustain_s |
0.5 | how long acceleration must be sustained |
heading_min_speed_ms |
3 | Cornering G undefined below this speed |
IMU and mount calibration (Phase 8)
| parameter | value | meaning |
|---|---|---|
imu_cutoff_hz |
5 | zero-phase Butterworth (order 4) cutoff on accelerometer and gyro |
imu_hampel_window |
11 | rolling-median window for clipped-sample replacement |
imu_hampel_nsigma |
None (disabled) | statistical Hampel threshold; None = disabled (Phase 8 finding) |
imu_replace_clipped |
true | replace samples at the ±8 g rail by the rolling median |
imu_decimate_hz |
50 | storage rate of filtered IMU channels |
sustained_window_s |
1 | window for sustained lateral values |
sustained_window_long_s |
0.5 | window for sustained longitudinal values (proof study) |
sustained_threshold_g |
1 | threshold for duration-above statistics |
calib_lowpass_hz |
1 | low-pass used for the mount calibration fit |
calib_long_min_ms2 |
2 | longitudinal window threshold (plan §14.6; rarely met on ovals) |
calib_lat_max_ms2 |
1 | lateral limit for straight-line windows |
calib_window_min_s |
1 | minimum straight-line window length |
calib_corr_good |
0.8 | lateral-axis correlation for confidence 1.0 |
calib_corr_ok |
0.6 | lateral-axis correlation for confidence 0.7 |
handedness_min_lat_ms2 |
2 | Cornering G magnitude that counts as a corner for the handedness test |
handedness_agreement_min |
0.95 | required lateral sign agreement |
lean_lowpass_hz |
0.5 | low-pass on the lean estimate used for the gyro lateral |
complementary_tau_s |
15 | complementary-filter time constant for the attitude estimate |
imu_agreement_rms_max_g |
0.15 | plan §14.7 item 3 threshold; superseded by the net-rotation identity (Amendment A2), reported not enforced |
stationary_up_spread_max_deg |
3 | stationary runs disagreeing by more use the pre-race run only |
mount_drift_max_deg |
10 | lap-mean up-vector drift that flags the mount as moved |
imu_agreement_bias_max_g |
0.05 | plan §14.7 bias threshold; reported, not enforced |
Centerline and projection (Phase 7)
| parameter | value | meaning |
|---|---|---|
centerline_bins |
180 | angular bins for the median polyline |
centerline_smooth_rms_m |
1 | target RMS residual of the smoothed centerline |
centerline_max_harmonics |
32 | maximum Fourier harmonics of the closed curve |
centerline_ds_m |
1 | grid spacing of s |
proj_sigma_s_m |
3 | expected along-track GPS jitter |
proj_back_tol_m |
2 | allowed backward move in progress |
proj_n_max_m |
30 | maximum lateral distance from the centerline |
proj_lost_limit |
5 | consecutive lost samples that invalidate a lap |
proj_min_window_m |
5 | minimum search window |
Segments (Phase 10)
| parameter | value | meaning |
|---|---|---|
segment_curvature_smooth_m |
10 | smoothing length of |
segment_min_length_m |
45 | minimum segment length after cleanup |
segment_user_min_length_m |
20 | shortest corner or straight a racer can draw in the segment editor |
segment_blob_m |
30 | corner runs shorter than this are ripple |
segment_corner_fraction |
0.35 | corner where |
segment_center_fraction |
0.15 | corner centre = ± this fraction of the arc |
Lap delta (Phase 7)
| parameter | value | meaning |
|---|---|---|
delta_invariant_tol_s |
0.001 | tolerance of the endpoint invariant |
delta_display_smooth_m |
10 | display smoothing window for Δt |
delta_low_confidence |
0.5 | projection confidence below which a grid point is low-confidence |