Smart City Road Network Optimization using Real-time GNSS/GIS Data Assimilation

Project Abstract

This study presents a novel framework for optimizing urban road networks through real-time GNSS and GIS data assimilation to enhance traffic efficiency, reduce travel times, and improve road safety in smart city environments. The research integrates heterogeneous data streams—GNSS traces from vehicles, crowd-sourced GPS data, high-resolution GIS layers, traffic sensor measurements, and municipal road network schemas—within a unified data fusion architecture based on state-space modeling and Bayesian inference. A multi-resolution data assimilation pipeline is developed to reconcile spatially and temporally diverse inputs, enabling accurate real-time estimation of traffic states, incident detection, and adaptive routing recommendations. The core contribution lies in the design of a real-time traffic state estimator that combines GNSS-based speed and position information with GIS-informed road geometry, lane-level constraints, and dynamic road attributes such as incidents, weather impacts, and construction zones. The estimator leverages a Kalman filter–inspired framework augmented with particle filter dynamics to handle nonlinearities and multimodal traffic patterns, ensuring robustness against sample sparsity and GNSS multipath effects. A novel map-matching algorithm with probabilistic confidence scoring maps raw GNSS trajectories to the most probable road segments, while accounting for urban canyon effects and variable signal quality. The research also introduces a scalable data assimilation engine that operates in near real-time on edge–cloud architectures, optimizing computational load through hierarchical data fusion and adaptive update rates. To validate the framework, synthetic and real-world datasets from a megacity environment are used to benchmark performance against baseline GPS-only and loop-detector–based approaches. Evaluation metrics include accuracy of traffic state estimation, incident detection latency, route optimality gains, and network-wide travel time reductions under varying demand scenarios and incidents. Results demonstrate substantial improvements in travel time reliability, with decreases in mean absolute percentage error for travel time predictions and faster detection of congestion onset. The study further analyzes the impact of data quality, penetration rates, and sensor heterogeneity on system performance, offering guidelines for data governance, privacy preservation, and data fusion parameterization. The practical implications include enhanced adaptive signal control, dynamic tolling strategies, and route guidance that collectively reduce congestion, lower emissions, and improve urban mobility resilience. A modular implementation blueprint is provided to facilitate replication and deployment in other urban contexts, highlighting considerations for interoperability with existing city ICT infrastructure, standards compliance, and stakeholder collaboration. limitations related to data availability, privacy constraints, and computational resource requirements are discussed, along with proposed mitigation strategies. The research concludes with a roadmap for future work focusing on integration with predictive-demand models, incorporation of connected and autonomous vehicle data, and the extension to multi-modal urban networks.

Project Overview

What This Project Is About

---

A straightforward, practical look at how city road networks can be improved by combining live GPS data with map information. The project explores how real-time location data from vehicles and sensors can be used with geographic maps to understand traffic patterns and suggest faster, safer routes.

---

---

The Problem It Addresses

---

Urban traffic often becomes congested due to outdated information, poor routing choices, and sudden incidents. This project tackles the gap between real-world traffic movement and the maps and data systems that guide drivers, aiming to reduce delays and improve road safety.

---

---

Objectives of the Project

---

1. Explain how real-time GNSS and GIS data work together to model road networks.
2. Demonstrate a simple method to detect traffic bottlenecks and incidents.
3. Develop a basic framework for suggesting alternative routes that save time or fuel.
4. Assess the usefulness of the approach in a small urban area.

---

---

What You Will Do Step by Step

---

1. Study key terms: GNSS, GIS, and data assimilation explained in plain language.
2. Collect real-time GPS data and map data for a chosen city area.
3. Merge data to build a dynamic road network model showing current traffic.
4. Identify patterns and potential route improvements using simple analysis.
5. Test a few routes and compare travel times under different conditions.
6. Evaluate limitations and propose practical improvements.

---

---

Expected Outcome

A clear, easy-to-use framework that shows how real-time data can improve route choices and reduce congestion, with a preliminary evaluation of benefits and limitations for city planners and transport agencies.