Empirical Evaluation of Obstacle Range Estimation Techniques in Diverse Environments using mmWave Radar

FMCW radars of the mmWave category possess a distinctive capability to ascertain the distance, speed, and angle of obstacles in their line of sight. These radars can identify obstructions at a maximum range of 350 to 400 meters. Range estimation is a critical step for the precise assessment of other metrics related to obstacles. An accurate range estimate is difficult in fluctuating environmental conditions or low/insufficient illumination. However, increasing the radar’s maximum detection range elevates environmental noise inference, hence compromising precise obstacle range estimates. Thus, this work provides a comparative examination of obstacle range estimates utilizing several CFAR detection techniques. The CFAR algorithm achieves obstacle detection by utilizing a threshold based on predicted background noise levels. The measurements are carried out under three realistic conditions, taking single and multiple obstacles that exist in these scenarios. The range of detected obstacles is compared with the utility provided by the TI. The analysis provided demonstrates the precision of each category of the CFAR algorithm. Thus, it is concluded that the order statistic CFAR demonstrates superior performance across all scenarios.