Comparative Analysis of Traffic Speed Data on Two-Lane Rural Road Segments: Radar Technology, Google Maps API, and License Plate Recognition

This paper presents a comparative analysis of traffic speed data on two-lane rural road segments using three different methods: radar technology measuring spot speeds, Google Maps API, and License Plate Recognition (LPR) providing space-mean speed estimates. The study employs a two-stage validation approach with LPR as the reference method, followed by direct comparison between Google Maps and radar speed data. Analysis of five road segments with varying geometry, traffic volumes, mobile signal quality, and radar placement reveals that no single method is universally superior. The findings show that Google Maps API offers a scalable, cost-effective solution with reasonable performance (MAE 5.5-8.8 km/h) under optimal conditions, but becomes unreliable in areas with weak mobile signal coverage and tends to smooth speed variations. When properly positioned, radar measurements provide high-resolution speed data with sensitivity to traffic changes (MAE 3.7-9.8 km/h against LPR), but their point-based nature creates significant dependency on sensor placement. Both methods exhibit reduced accuracy during low-speed conditions, as indicated by elevated MAPE values. Mobile network signal quality emerges as critical for Google Maps reliability, while road segment geometry and sensor positioning are paramount for radar reliability.