Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Background Subtraction with Drift Correction for Bistatic Radar Reflectivity Measurements

Alexander Ihlow, Marius Schmidt, Carsten Andrich, Reiner S. Thomä

TL;DR

The paper addresses drift-induced coherence loss in background-subtracted bistatic radar reflectivity measurements in anechoic chambers by proposing a parametric drift-correction model with three real-valued parameters that jointly adjust amplitude and phase. The correction uses a time-domain residual minimization solved with Newton-Conjugate-Gradient, leveraging analytic derivatives, and is validated on long-term static, multi-position, and sphere measurements, achieving up to 40 dB improvement in suppressing direct-path leakage. The results demonstrate that the method enhances reflectivity extraction in the monostatic/bistatic regime, though it is not applicable in forward-scattering regions where direct path and target signals merge. The approach has implications for high-frequency radar research and ISAC applications, improving measurement fidelity in complex chambers; future work includes interpolating parameters across angular regions.

Abstract

Fundamental research on bistatic radar reflectivity is highly relevant, e.g., to the upcoming mobile communication standard 6G, which includes integrated sensing and communication (ISAC). We introduce a model for correcting instrumentation drift during bistatic radar measurements in anechoic chambers. Usually, background subtraction is applied with the goal to yield the target reflection signal as best as possible while coherently subtracting all signals which were present in both the foreground and background measurement. However, even slight incoherences between the foreground and background measurement process deteriorate the result. We analyze these effects in real measurements in the frequency range 2-18 GHz, taken with the Bistatic Radar (BIRA) measurement facility at TU Ilmenau. Applying our proposed drift correction model, we demonstrate up to 40 dB improvement for the removal of direct line-of-sight antenna crosstalk over the state of the art.

Background Subtraction with Drift Correction for Bistatic Radar Reflectivity Measurements

TL;DR

The paper addresses drift-induced coherence loss in background-subtracted bistatic radar reflectivity measurements in anechoic chambers by proposing a parametric drift-correction model with three real-valued parameters that jointly adjust amplitude and phase. The correction uses a time-domain residual minimization solved with Newton-Conjugate-Gradient, leveraging analytic derivatives, and is validated on long-term static, multi-position, and sphere measurements, achieving up to 40 dB improvement in suppressing direct-path leakage. The results demonstrate that the method enhances reflectivity extraction in the monostatic/bistatic regime, though it is not applicable in forward-scattering regions where direct path and target signals merge. The approach has implications for high-frequency radar research and ISAC applications, improving measurement fidelity in complex chambers; future work includes interpolating parameters across angular regions.

Abstract

Fundamental research on bistatic radar reflectivity is highly relevant, e.g., to the upcoming mobile communication standard 6G, which includes integrated sensing and communication (ISAC). We introduce a model for correcting instrumentation drift during bistatic radar measurements in anechoic chambers. Usually, background subtraction is applied with the goal to yield the target reflection signal as best as possible while coherently subtracting all signals which were present in both the foreground and background measurement. However, even slight incoherences between the foreground and background measurement process deteriorate the result. We analyze these effects in real measurements in the frequency range 2-18 GHz, taken with the Bistatic Radar (BIRA) measurement facility at TU Ilmenau. Applying our proposed drift correction model, we demonstrate up to 40 dB improvement for the removal of direct line-of-sight antenna crosstalk over the state of the art.
Paper Structure (6 sections, 6 equations, 10 figures, 3 tables)

This paper contains 6 sections, 6 equations, 10 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Drift Measurement
  3. Correction Model
  4. Multi-Position Measurement
  5. Application onto Sphere Measurements
  6. Conclusion

Figures (10)

  • Figure 1: Phase deviations in static long-term measurement.
  • Figure 3: Impulse responses of static long-term measurement. The set of curves correspond to \ref{['fig:longterm_phase']} and \ref{['fig:longterm_mag']} (cf. the annotation of measurement time in \ref{['fig:longterm_phase']}).
  • Figure 4: Impulse responses after conventional background subtraction (without correction): The first measurement is treated as "background" and subtracted from all other measurements. The set of curves correspond to \ref{['fig:longterm_phase']} and \ref{['fig:longterm_mag']} (cf. the annotation of measurement time in \ref{['fig:longterm_phase']}).
  • Figure 5: Impulse responses after background subtraction with correction: The first measurement is treated as "background" and subtracted from all other measurements. The model according to \ref{['eq:model']} is applied, correcting phase and amplitude deviations. This results in an improvement of up to $40$ dB regarding peak subtraction.
  • Figure 6: Performance of background subtraction over all static measurement runs (cf. \ref{['tab:static_measurement_parameters']}). The first measurement is treated as "background". The proposed correction model gains up to $40$ dB peak improvement compared to conventional background subtraction.
  • ...and 5 more figures