Advanced Forecasting and Engineering Design Solutions to Improve Traffic Flow on Major Road Networks

Abstract

This study examines the integration of advanced forecasting techniques and engineering design solutions to enhance traffic flow on Interstate 95 (I-95) between Miami and West Palm Beach, Florida. Using traffic data from 2015–2023, the research compared a baseline SARIMA model (MAPE 12.8%) against an LSTM model with real-time inputs (MAPE 9.6%), achieving a 25% accuracy gain. Engineering interventions—adaptive traffic signal control (ATSC) and ramp metering—reduced peak-hour travel times by 17% (78 minutes), while a combined framework cut times by 22% (72 minutes), emissions by 19%, and crash risk by 25%. Annual savings could reach $100 million in congestion costs and 20,000 tons CO₂. Findings highlight the synergy of predictive analytics and infrastructure optimization, though rural data gaps and costs pose challenges. Recommendations include expanding sensor networks, prioritizing ATSC deployments, and fostering regional coordination. This framework offers a scalable model for U.S. road networks