Environment-Aware Learning of Smooth GNSS Covariance Dynamics for Autonomous Racing
Y. Deemo Chen, Arion Zimmermann, Thomas Berrueta, Soon-Jo Chung
AI summary
Problem
High-speed autonomous systems require GNSS uncertainty estimates that adapt to environmental changes while remaining temporally smooth to prevent controller instability, but existing methods fail to jointly model environmental adaptation and covariance dynamics.
Approach
A neural network with spatial attention predicts process noise from environmental features, which is then propagated through a spectrally constrained Lyapunov differential equation to guarantee stable and smooth covariance evolution by design.
Key results
- Novel architecture unifying environmental adaptation and tunable smoothness
- Formal proof of exponential stability and convergence for covariance dynamics
- Improved localization accuracy and smoother uncertainty in GNSS-degraded racing environments
- Data-efficient parallelizable implementation enabling real-time deployment
Why it matters
It enables reliable, high-speed autonomous navigation in GNSS-challenged environments by preventing control destabilization caused by erratic uncertainty estimates.
Abstract
Ensuring accurate and stable state estimation is a challenging task crucial to safety-critical domains such as high-speed autonomous racing, where measurement uncertainty must be both adaptive to the environment and temporally smooth for control. In this work, we develop a learning- based framework, LACE, capable of directly modeling the temporal dynamics of GNSS measurement covariance. We model the covariance evolution as an exponentially stable dynamical system where a deep neural network (DNN) learns to predict the system’s process noise from environmental features through an attention mechanism. By using contraction- based stability and systematically imposing spectral constraints, we formally provide guarantees of exponential stability and smoothness for the resulting covariance dynamics. We validate our approach on an AV-24 autonomous racecar, demonstrating improved localization performance and smoother covariance estimates in challenging, GNSS-degraded environments. Our results highlight the promise of dynamically modeling the perceived uncertainty in state estimation problems that are tightly coupled with control sensitivity.