Matched-Filter Radar Signal Processing With Image Enhancement For Long-Range Target Detection, Coverage Analysis, and Plot-Based Performance Evaluation At 1 Km Range

Abstract

Reliable long-range target detection in automotive radar systems remains challenging under low signal-to-noise ratio (SNR), clutter, and interference conditions. This study proposes a matched-filter-based radar signal processing model for long-range target detection at a range of 1 km using a 77 GHz FMCW radar sensor. The proposed model incorporates waveform generation, target channel modeling, additive white Gaussian noise (AWGN) addition, pulse compression using matched filters, Cell-Averaging Constant False Alarm Rate (CA-CFAR) detection, image-processing-based false alarm suppression, and radar coverage analysis within a unified processing architecture. The matched-filter implementation achieves a time-bandwidth product of 1.5 x 10⁵ leading to pulse compression gain of about 51.8 dB. It enhances the processing SNR from -22.5 dB before processing to an SNR of 24.3 dB after processing at 1 km target range. The created RDM image is analyzed using CA-CFAR detector with 0.97 detection probability at 10⁻⁶ false alarm probability in a noisy environment. For the purpose of enhancing the detection capability, a two-stage enhancement framework consisting of Otsu re-thresholding, morphological processes, and blob filtering is proposed. Using the proposed enhancement approach, there is a significant reduction in false alarm rate by up to 99% without compromising target detection capability. The range coverage analysis indicates a maximum detection distance of 1620 m. Extensive graphical evaluations validate the robustness and effectiveness of the proposed framework for automotive radar applications operating in low-SNR and cluttered environments.