Road Transit Time Forecast Correction Using Cellular Localization Data

The patent merges multiple data sources including cellular network localization data with other traffic information sources to compensate for the imprecision of individual cellular localization. By combining data from multiple terminals and cross-referencing with road network geometry and historical patterns, the system achieves acceptable measurement precision while maintaining capillary coverage.

The system uses successive localization copies of the same mobile terminal to track vehicle movement. Instead of relying on a single precise measurement, it collects multiple localization points over time and uses these copies to infer traffic flow patterns, thereby overcoming the limitation of individual measurement precision.

The system implements feedback mechanisms by comparing estimated transit times with actual observed times and adjusting the interpretation of localization data accordingly. When anomalies are detected (such as unusually long stops that may indicate phone call behavior), the system uses feedback from multiple data points to correct for these deviations and maintain reliable traffic estimates.

The patent collects excessive localization data points beyond what would be minimally required, then uses statistical filtering to extract reliable traffic information. By gathering more data than strictly necessary and applying filtering algorithms, the system can tolerate some anomalous data from phone-using drivers while still achieving accurate traffic flow monitoring.