A Reconfigurable Framework for Vehicle Localization in Urban Areas

Viana, Kerman; Zubizarreta, Asier; Diez, Mikel · 2022 · Crossref

DOI: 10.3390/s22072595

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 addresses the challenge of maintaining accurate vehicle localization in dense urban environments where sensor failures, such as GPS signal interruption or sensor occlusion, are frequent. Traditional autonomous vehicle systems often respond to detected failures by immediately stopping the vehicle, which limits operational continuity. The authors propose a reconfigurable localization framework that allows vehicles to continue driving in a degraded mode by dynamically switching to alternative positioning strategies until a critical failure necessitates a stop. The framework is designed to handle temporary failures and reset accumulated drift when sensors recover. The proposed framework utilizes a hierarchical structure with three localization levels. Level 1 provides the most accurate global positioning using GPS/INS fusion. Level 2 combines relative positioning estimates with digital map data via road-matching algorithms, supplemented by junction detection to correct longitudinal drift. Level 3 relies on relative positioning using visual odometry, LiDAR odometry, inertial measurement units, and wheel odometry, fused via an Extended Kalman Filter (EKF). An error-detection block monitors these levels; for Level 3, it employs a conflict evaluation algorithm that compares sensor outputs to calculate dynamic reliability values, adjusting the EKF’s measurement noise covariance to downweight faulty sensors. A decision block manages reconfiguration, switching between levels based on error states and enforcing time thresholds to prevent excessive drift in degraded modes. The framework was validated using the CARLA simulation environment across various urban scenarios with simulated sensor errors. Results demonstrated that the system maintained proper localization during sensor failures by reconfiguring to alternative strategies. The reconfiguration module successfully reset accumulated drift in alternative algorithms when temporary failures resolved, thereby bounding the mean error. The vehicle continued operation in degraded modes (Level 2 or Level 3) only stopping when the system reached a fully degraded state or exceeded safety time thresholds. The significance of this work lies in its ability to enhance the robustness and availability of autonomous vehicles in urban settings. By enabling continued operation during temporary sensor degradation rather than immediate cessation, the framework improves safety and utility. The approach highlights the importance of integrating fault detection with dynamic reconfiguration strategies, offering a practical solution for managing the inherent unreliability of sensor data in complex environments.

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.

• situational awareness