RF Multi-Signal Analysis Report

ITAR Screen (Upload + Outputs)

ITAR Screen Result = CLEAR

Basis: The ZIP contains commercial/civil RF examples (NOAA weather, VHF airband AM, APRS, AIS, DMR, POCSAG, LoRa, GSM, BLE, Wi-Fi) with no obvious ITAR/USML indicators (e.g., weapon systems, fire-control radar, seeker telemetry, military datalinks, classified/controlled parameters). File naming and the included index CSV appear training/demo-oriented.

1. Executive Summary

2. Data Characteristics

Detected data type (hardwired): A) Raw IQ / time-series (complex I/Q)

• Container: WAV (PCM_16), 2 channels, sample rate 48,000 Sa/s
• Duration per file: 30 s (matches CSV)
• Implied complex bandwidth: ~48 kHz (limited)

• Clipping/overload check: clip fraction ≈ 0 across files (no near-full-scale saturation observed)
• Calibration: No absolute dBm scaling available; power/SNR are relative

3. Detected Signals / Events (per recording)

All recordings appear signal-present nearly continuously over the 30 s capture (no long idle periods detectable with energy gating), so “duty cycle” in the usual sense can’t be estimated reliably here.

Measured occupied bandwidth (OBW) below is a thresholded PSD estimate (relative), useful for comparing files within this dataset.

01_NOAA_NBFM_162_550MHz.wav CF 162.550 MHz

• Measured OBW: ~12.0 kHz
• Envelope: near-constant → consistent with FM/constant-envelope
• Classification: Narrowband FM broadcast-like (NOAA WX plausible by CF)
• Confidence: 85% (High)

02_VHF_AIR_AM_125_700MHz.wav CF 125.700 MHz

• Measured OBW: ~31.2 kHz
• Envelope: high variation (AM-like envelope)
• Classification: AM voice-like (VHF airband plausible by CF)
• Confidence: 90% (High)

03_APRS_AFSK_144_390MHz.wav CF 144.390 MHz

• Measured OBW: ~1.68 kHz
• Behavior: narrowband, tone/FSK-like consistent with AFSK packet
• Classification: APRS/AFSK-style narrowband data
• Confidence: 88% (High)

04_AIS_MARITIME_162_025MHz.wav CF 162.025 MHz

• Measured OBW: ~6.6 kHz
• Behavior: narrowband constant-envelope FSK-like
• Classification: AIS-like narrowband data
• Confidence: 85% (High)

05_DMR_451MHz.wav CF 451.000 MHz

• Measured OBW: ~6.56 kHz
• Behavior: constant-envelope digital voice/data-like
• Classification: DMR-like narrowband digital (UHF LMR plausible)
• Confidence: 80% (Moderate–High)

06_POCSAG_PAGER_929_612MHz.wav CF 929.612 MHz

• Measured OBW: ~6.67 kHz
• Behavior: narrowband FSK-like
• Classification: Paging FSK / POCSAG-like
• Confidence: 82% (Moderate–High)

07_LORA_915MHz.wav CF 915.000 MHz

• Measured OBW: ~4.02 kHz
• Behavior: frequency-sweeping / chirp-ish indicator (within capture constraints)
• Classification: LoRa CSS-like (constrained)
• Confidence: 70% (Moderate)

08_GSM_BURST_935MHz.wav CF 935.000 MHz

• Measured OBW: extremely low by this estimator (not trustworthy here)
• Behavior: not like realistic full-rate GSM at 48 kHz BW; likely decimated/synthetic
• Classification: GSM/TDMA-burst label plausible by CF; waveform evidence insufficient
• Confidence: 45% (Insufficient–Moderate)

09_BLUETOOTH_LE_2_402GHz.wav CF 2402.000 MHz

• Measured OBW: ~5.64 kHz
• Behavior: NB-ish in this representation; not consistent with true 1–2 MHz BLE unless heavily decimated
• Classification: BLE advertising plausible by CF; evidence limited
• Confidence: 50% (Insufficient–Moderate)

10_WIFI_BEACON_2_437GHz.wav CF 2437.000 MHz

• Measured OBW: ~48.0 kHz (fills available complex BW)
• Behavior: wide/complex within 48 kHz, but insufficient BW to validate OFDM Wi-Fi
• Classification: Wi-Fi beacon/OFDM label plausible by CF; not verifiable here
• Confidence: 40% (Insufficient)

4. Classification Summary Table

*OBW estimator becomes unreliable when the capture BW is far smaller than the true signal BW or when the signal violates estimator assumptions.

5. Limitations & Assumptions

• Bandwidth/sample-rate constraint: 48 kSa/s complex limits you to ~48 kHz RF bandwidth. Fine for narrowband services, but insufficient to confirm Wi-Fi / GSM / BLE as real over-the-air captures.
• No absolute calibration: Power/SNR values are relative (no dBm reference, unknown gain/AGC).
• No ground-truth capture context: Antenna, gain, filtering, RF front-end conditions not provided; restricts confidence.
• Propagation effects: Outdoor ground environment can include multipath/fading; insufficient metadata to separate channel effects from modulation traits.

Under optimal conditions, classification can exceed 95%; real confidence depends on sample rate/bandwidth, front-end dynamic range, gain/AGC, clipping, noise figure, antenna & placement/polarization, channel conditions, clock error, frequency accuracy, filtering, compression, capture completeness, and ground truth.

6. Recommendations

1. For unknown real-world emitters: capture wider IQ bandwidth that matches suspected service:
   • Wi-Fi: ≥20 MHz (prefer 40/80 if feasible)
   • BLE: ≥2–4 MHz recommended
   • GSM: ≥1–2 MHz (more if scanning multiple carriers)
2. Record and share metadata: SDR model, gain/AGC, antenna type/height, front-end filters, approximate location, and whether captures are baseband IQ or post-demod audio.
3. If interference hunting is the goal: collect multiple captures at different times/positions and add direction-finding (even crude antenna rotation logs help).
4. If you want reclassification “from scratch” (ignore CSV labels), provide raw wideband IQ for the unknown band(s) and I’ll run a segmentation-first workflow (signals → features → conservative class hypotheses).