Microwave Radar Vehicle Detection Performance at Railroad Grade Crossings with Quad Gates in Normal and Adverse Weather Conditions

This paper presents a comprehensive all-weather evaluation of a dual microwave radar vehicle detection system at a highway–rail grade crossing equipped with four quadrant (quad) gates. Weather conditions ranged from favorable (normal) to adverse and included rain (light and heavy), snow (light and heavy), fog, and wind. Initial assessment of the detection performance in adverse weather showed higher than acceptable error rates mostly in heavy rain (false calls > 30%) and heavy snow (systemwide missed calls > 11%). However, after detailed feedback was provided to the vendor and the system parameters were remodified, a subsequent reevaluation was made by using more than 110,000 radar activations and close to 450 h of data. Final results showed that the remodified setup not only reduced the frequency of errors in heavy rain and heavy snow conditions but also maintained a desirable performance in the other weather conditions (including light rain and light snow). After remodification, there were no systemwide missed calls, stuck-on calls, or dropped calls, with less than 1% detection errors in good weather and up to 2.6% during specific periods of adverse weather. This paper describes the process followed to achieve this performance in the field. To ensure reliability, before the system is set as the primary gate control mechanism, thorough evaluation and monitoring of similar systems at other locations is recommended, particularly at wider crossings and longer crossings.