Two microwave-based systems for vehicle detection (by Wavetronix and MS SEDCO) were evaluated at stop bar and advance zones of a signalised intersection under three adverse weather conditions: (1) wind, (2) snow-covered roadway, and (3) rain. Weather effects were very different for the two systems both in terms of the type of condition that could affect performance and in the magnitude of those effects. For Wavetronix, wind had significant effects on the advance zone by increasing false calls to over 50%, but it did not affect the stop bar zones. On the other hand, false calls in snow significantly increased to more than 40% in the stop bar zones and to about 30% in the advance zone. Snow also increased missed and stuck-on calls but in lower proportion than the false calls. Rain also affected the detection at stop bar zones, but all error types were below 8%, and it did not affect the advance zone. For Intersector, weather effects were less pronounced both at the stop bar and advance zones. Snow increased false calls to a range of about 4% to 8% compared to 1.65% to about 4% in normal weather. In addition, rain increased stuck-on calls to a range of 2.7% to 6.35% at the stop bar zones and increased missed calls at advance zones to 3.44%. Wind had no significant effects at stop bar or advance zones. In particular for the rain data, the intensity of the precipitation seemed to be related to the degree of performance degradation. In datasets with higher precipitation per unit of time, higher false calls were observed at Wavetronix stop bar zones, and a higher frequency of missed calls was observed at the Intersector advance zone. Findings from this evaluation can provide valuable information to users and manufacturers of these products regarding expected performance under adverse weather conditions at locations with similar mountings and settings, as well as insight about potential solutions to preventing negative effects in such scenarios. (Author/publisher)
Abstract