UAV-based Intelligent Information Systems on Winter Road Safety for Autonomous Vehicles

TL;DR

The paper addresses the challenge of autonomous vehicle operation in winter conditions by proposing a UAV-based intelligent winter road information system that provides up-to-date road geometry and surface condition data to AVs. It integrates a dual-perspective data collection platform, combining a ground vehicle-mounted data collector and an unmanned aerial sensor suite, and develops methods for road width estimation from 3D point clouds and snow-heap assessment from road surface profiles, complemented by DEM generation and photogrammetry workflows. Experimental results show the UAV-based approach can estimate road width and surface conditions, but segmentation under snow is challenging, while LiDAR intensity from the autonomous car perspective can aid lane detection when markings are obscured. The work demonstrates the practical potential of UAV-assisted ITS for winter road safety and highlights the need for improved segmentation models and robust DEM/edge definitions to support real-time AV decision-making.

Abstract

As autonomous vehicles continue to revolutionize transportation, addressing challenges posed by adverse weather conditions, particularly during winter, becomes paramount for ensuring safe and efficient operations. One of the most important aspects of a road safety inspection during adverse weather is when a limited lane width can reduce the capacity of the road and raise the risk of serious accidents involving autonomous vehicles. In this research, a method for improving driving challenges on roads in winter conditions, with a model that segments and estimates the width of the road from the perspectives of Uncrewed aerial vehicles and autonomous vehicles. The proposed approach in this article is needed to empower self-driving cars with up-to-date and accurate insights, enhancing their adaptability and decision-making capabilities in winter landscapes.

Figures (12)

• Figure 1: (LHS) Sensor setup mounted in front of the vehicle with sensor labels. (RHS) Data flow in the framework. Raw data is read from the sensor, converted to SI-units and saved to tmpfs by sensor writer. Data is then read from temporary file system using sensor reader object and processed for future use. Processed data is saved back in the tmpfs and can be read by next part of the processing pipeline. Data in the tmpfs can be utilized for real time robot control or saved for offline use by data saver.
• Figure 2: UAV Sensor system architecture
• Figure 3: Road Cross Section Design Elements in Pohjantie
• Figure 4: Roundabout and Road Cross Section Design Elements in Ruskontie
• Figure 5: Pohjantie Highway 38.5m: (a) 3D Generated colored dense cloud, (b) 3D ROI Segmented Road surface, (c) 2D ROI UAV Ortho images