Dynamic traffic optimization systems ensure data privacy.

Abstract

The invention relates to a dynamic traffic optimization system combining linkage learning and fuzzy logic with data privacy preservation. The system comprises traffic IoT sensor nodes (10), area fog computing nodes (20), linkage learning aggregator (30), congestion prediction and mapping module (40), route optimization module (50) and smart city dashboard (60). Each area fog computing node (20) trains a local traffic prediction model (220) and computes the fuzzy congestion coefficient Ck using Mamdani-type fuzzy inference (230), while raw traffic data is kept local. The linkage learning aggregator (30) receives the local model parameter set θ'k and the fuzzy congestion coefficient Ck, determines the aggregate weight wk and computes the global model θglobal based on the fuzzy weighted linkage learning average. The congestion prediction and mapping module (40) uses the global model θglobal to generate congestion coefficients Ci.The route optimization module (50) updates the competition cost according to the congestion coefficients Ci, and applies the Dijkstra algorithm to determine the optimal route. This helps reduce raw data transmission, preserve data privacy, and support real-time route optimization.

Images