Uncertainty-Aware Autonomous Vehicles: Predicting the Road Ahead
Shireen Kudukkil Manchingal, Armand Amaritei, Mihir Gohad, Maryam Sultana, Julian Francisco Pieter Kooij, Fabio Cuzzolin, Andrew Bradley
AI summary
Problem
Autonomous vehicle perception systems frequently make overconfident predictions in rare or ambiguous scenarios, leading to unsafe decisions when encountering out-of-distribution data or edge cases.
Approach
The authors embed a Random-Set Neural Network classifier into a ROS-based control pipeline to predict upcoming road layouts while explicitly quantifying uncertainty, which dynamically scales vehicle speed in real time.
Key results
- RS-NN achieved 75.54% accuracy on a 7-class road layout dataset, outperforming CNNs and Bayesian networks
- RS-NN entropy distributions clearly separate regular from ambiguous test data
- Real-time integration dynamically scaled vehicle speed based on classification entropy
- Introduced the TrackDrive Direction dataset with 1,187 images for layout classification and ambiguity testing
Why it matters
This framework offers a practical, real-time solution for safer high-speed autonomous driving by enabling vehicles to proactively adjust behavior when perception confidence drops.
Abstract
Autonomous Vehicle (AV) perception systems have advanced rapidly in recent years, providing vehicles with the ability to accurately interpret their environment. Perception systems remain susceptible to errors caused by overly-confident predictions in the case of rare events or out-of-sample data. This study equips an autonomous vehicle with the ability to ’know when it is uncertain’, using an uncertainty-aware image classifier as part of the AV software stack. Specifically, the study exploits the ability of Random-Set Neural Networks (RS-NNs) to explicitly quantify prediction uncertainty. Unlike traditional CNNs or Bayesian methods, RS-NNs predict belief functions over sets of classes, allowing the system to identify and signal uncertainty clearly in novel or ambiguous scenarios. The system is tested in a real-world autonomous racing vehicle software stack, with the RS-NN classifying the layout of the road ahead and providing the associated uncertainty of the prediction. Performance of the RS-NN under a range of road conditions is compared against traditional CNN and Bayesian neural networks, with the RS-NN achieving significantly higher accuracy and superior uncertainty calibration. This integration of RS-NNs into Robot Operating System (ROS)-based vehicle control pipeline demonstrates that predictive uncertainty can dynamically modulate vehicle speed, maintaining high-speed performance under confident predictions while proactively im- proving safety through speed reductions in uncertain scenarios. These results demonstrate the potential of uncertainty-aware neural networks - in particular RS-NNs - as a practical solution for safer and more robust autonomous driving.