iNode adaptive outlier filtering algorithm

This white paper introduces SMATS’ iNode Adaptive Filtering Algorithm, designed to clean noisy travel time data collected from Bluetooth and Wi-Fi MAC address detections. The algorithm effectively identifies and removes outliers in both arterial and highway contexts, adapting to various sample sizes and abrupt traffic changes caused by incidents or fluctuations.

Through multiple real-world deployment examples—including Toronto highways, Minneapolis rush hours, signalized arterials, and port entry delays in Los Angeles—the algorithm consistently filtered inaccurate data, isolating true traffic trends. Its performance in scenarios with parallel traffic streams and low data volumes highlights its robustness, enabling agencies to trust the integrity of their real-time traffic analytics and improve operational decision-making.