Clutter-Aware Weak-Target Detection in Passive-Radar Range– Doppler Data Using Calibrated Recursive Complex-Valued Convolutive ICA

Abstract Weak-target detection in complex range–Doppler (RD) observations is challenging in heterogeneous clutter, particularly when raw reference and surveillance channels are unavailable. This study examines a clutter-aware detector combining frame-sequential recursive complex-valued convolutive independent component analysis (ICA) with interquartile-range (IQR) score calibration and maximum-rank fusion with a local CFAR-like baseline. Score orientation uses labeled calibration observations, whereas target-absent (H0) observations provide robust normalization and threshold estimates. Real-data evaluation uses only the processed complex RD measurements of the publicly available LSS-PR-1.0 dataset. In a pooled evaluation, the proposed H0-CG-OR-CVCICA detector achieved an area under the receiver operating characteristic curve (AUC) of 0.719, compared with 0.578 for the evaluated CA-CFAR-like baseline. Across 30 repeated file/group-level calibration/test splits, mean AUC was 0.714 ± 0.029 versus 0.570 ± 0.015, and mean detection probability at nominal false-alarm probability 0.01 was 0.260 ± 0.031 versus 0.132 ± 0.014. The proposed method had an empirical false-alarm probability of 0.015 and a reported processing time of 1.320 ms per frame in the tested Python implementation. Separate synthetic in-phase/quadrature (IQ) experiments were used exclusively to examine extensive cancellation algorithm (ECA)-type baselines, not for direct comparison on LSS-PR-1.0. Performance varied across target categories, and the two independent H0 clutter recordings limit conclusions about generalization to new recording conditions. The results support the usefulness of calibrated component-derived scores under the evaluated offline RD-domain protocol, without establishing end-to-end streaming operation or superiority over untested CFAR variants.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-17
DOI
https://doi.org/10.21203/rs.3.rs-11431862/v1
Primary Topic
Radar Systems and Signal Processing
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Clutter-Aware Weak-Target Detection in Passive-Radar Range– Doppler Data Using Calibrated Recursive Complex-Valued Convolutive ICA

Hamzeh Ghahramani
1 links
Research Square
Radar Systems and Signal Processing
preprint

Clutter-Aware Weak-Target Detection in Passive-Radar Range– Doppler Data Using Calibrated Recursive Complex-Valued Convolutive ICA

Hamzeh Ghahramani
preprint en

Abstract

Abstract Weak-target detection in complex range–Doppler (RD) observations is challenging in heterogeneous clutter, particularly when raw reference and surveillance channels are unavailable. This study examines a clutter-aware detector combining frame-sequential recursive complex-valued convolutive independent component analysis (ICA) with interquartile-range (IQR) score calibration and maximum-rank fusion with a local CFAR-like baseline. Score orientation uses labeled calibration observations, whereas target-absent (H0) observations provide robust normalization and threshold estimates. Real-data evaluation uses only the processed complex RD measurements of the publicly available LSS-PR-1.0 dataset. In a pooled evaluation, the proposed H0-CG-OR-CVCICA detector achieved an area under the receiver operating characteristic curve (AUC) of 0.719, compared with 0.578 for the evaluated CA-CFAR-like baseline. Across 30 repeated file/group-level calibration/test splits, mean AUC was 0.714 ± 0.029 versus 0.570 ± 0.015, and mean detection probability at nominal false-alarm probability 0.01 was 0.260 ± 0.031 versus 0.132 ± 0.014. The proposed method had an empirical false-alarm probability of 0.015 and a reported processing time of 1.320 ms per frame in the tested Python implementation. Separate synthetic in-phase/quadrature (IQ) experiments were used exclusively to examine extensive cancellation algorithm (ECA)-type baselines, not for direct comparison on LSS-PR-1.0. Performance varied across target categories, and the two independent H0 clutter recordings limit conclusions about generalization to new recording conditions. The results support the usefulness of calibrated component-derived scores under the evaluated offline RD-domain protocol, without establishing end-to-end streaming operation or superiority over untested CFAR variants.

Research Square
Islamic Azad University Bandar Abbas (IR), Islamic Azad University, Bandar Anzali Branch (IR)
Radar Systems and Signal Processing
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon