Signal quality indexes for millimeter-wave radar: enhancing heart rate estimation via sub-range bin selection

This is the authors' abstract. We don't add key points for this paper.

The accuracy of millimeter-wave radar-based heart rate estimation is directly related to the selection of a range bin that maximizes heartbeat information. However, this task remains challenging, as interference often masks the heartbeat signal, and the target’s position within a single range bin further affects measurement reliability. Crucially, existing range bin selection methods do not adequately account for this intra-bin location dependency of signal quality or its susceptibility to various interference sources. To address these limitations, this study proposes a physical sub-range bin selection strategy. This approach is founded on the hypothesis that each range bin can be further subdivided into multiple sub-range bins. Signal quality varies spatially across these subdivisions, and small physical displacements of the radar can shift the target’s effective position from disturbed sub-regions to optimal sub-range bins. To implement this, a precision electric linear stage was utilized to perform controlled radar displacements, and a composite Signal Quality Index ( SQ I radar ) was developed to quantify diverse interference components and isolate sub-range bins with minimal disturbance. Specifically, the index integrates two novel indicators with three established metrics. Principal Component Analysis (PCA) was subsequently employed to mitigate redundancy and extract a minimally correlated subset of indicators for the final index construction. The results indicated that SQ I radar successfully selected a higher-quality sub-range bin in over 80% of the samples, reducing the mean relative error by 49% (from 0.095 to 0.048). In real-time experiments, when the SQ I radar ≥ 1.00, a relative error below 0.10 was achieved in 78% of the samples.