PROUD-Public road urban driverless test: Architecture and results

Broggi, Alberto; Cerri, Pietro; Debattisti, Stefano; Laghi, Maria Chiara; Medici, Paolo; Panciroli, Matteo; Prioletti, Antonio · 2014 · OpenAlex-citations

DOI: 10.1109/ivs.2014.6856478

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper presents the architecture and results of the PROUD 2013 (Public ROad Urban Driverless test), a demonstration of autonomous driving on open public roads in Parma, Italy. The study addresses the transition of autonomous vehicle testing from controlled environments to real-world traffic conditions, motivated by new regulations allowing such tests and the necessity to validate vehicle performance in complex, unstructured scenarios. The primary objective was to demonstrate the capability of the BRAiVE prototype to navigate mixed environments—including rural roads, highways, and urban centers—without human intervention, handling dynamic obstacles, traffic lights, roundabouts, and pedestrian crossings. The experimental setup utilized the BRAiVE vehicle, equipped with an enhanced sensor suite including IP67-rated laser scanners, stereo cameras, and a GNSS device with real-time kinematic correction for high-precision localization. The software architecture, developed under the OFAV project, is modular and scalable, featuring a strict separation between perception, planning, and control layers. Communication between components is managed by VisLab Bus (VLBus), a message-passing service. A central component, the World Perception Server (WPS), performs multilevel sensor fusion to create a unified perception map of obstacles, lanes, barriers, traffic lights, and road markings. The planning layer is hierarchical, comprising a Navigation layer for route generation using Open Street Map data, a Maneuver layer that selects driving modes and generates traversability maps, and a Control layer that executes path planning using precomputed trajectories and Model Predictive Control for steering and acceleration. The test was conducted on July 12, 2013, covering a 13 km route that included 10.6 km of highway, 2.4 km of urban/suburban roads, and 0.7 km of rural roads. The vehicle completed the entire journey autonomously in 18 minutes, achieving an average speed of 43.33 km/h and a maximum speed of 70 km/h. The system successfully navigated six roundabouts, three junctions, 15 crosswalks, and eight tunnels. Notably, the vehicle maintained autonomy even during a GNSS outage lasting 4 minutes and 22 seconds, relying on inertial sensors and visual perception. The results indicate that the architecture effectively managed common infrastructures and integrated with normal traffic, though the authors note that the specific path was previously tested and that significant challenges remain for full automation in unpredictable urban environments. The significance of this work lies in its demonstration of a robust, modular architecture capable of handling the complexities of public road driving. By successfully executing a mixed-environment route without human intervention, the study validates the feasibility of autonomous driving in real-world conditions and highlights the importance of robust perception algorithms and hierarchical planning. The authors conclude that while this test represents a substantial step toward full automation, further research is required to address unsolved challenges in dynamic and unpredictable traffic scenarios.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• driverless ads