Institution profile

Federal University of Santa Catarina

Academic institutionsouthamerica · br
Official website
Research library33linked papers
Opportunities0open roles
Selected work

Representative Papers

Position: Certifiable State Integrity in Cyber-Physical Systems -- Why Modular Sovereignty Solves the Plasticity-Stability Paradox

Jan 29, 2026

This work addresses the challenges faced by monolithic temporal foundation models in safety-critical cyber-physical systems, where catastrophic forgetting, oversmoothing of high-frequency fault signals, and lack of verifiability compromise lifecycle-wide state integrity. To overcome these limitations, the authors propose a modular sovereignty paradigm comprising a frozen library of condition-specific expert models, coupled with an uncertainty-aware hierarchical fusion mechanism (HYDRA) that rigorously disentangles aleatoric and epistemic uncertainties. This approach enables module-level auditability and validity guarantees, effectively resolving the plasticity–stability dilemma. By doing so, it provides a certifiable pathway aligned with functional safety standards such as ISO 26262, ensuring high robustness and state integrity under non-stationary operating conditions.

1 citations1 influentialRead paper

Closed-Loop Evaluation of Bird's-Eye-View Maps from Cross-View Transformers as Inputs to Behavior-Cloning Policies

Sep 05, 2026

In autonomous driving, Bird's-Eye View (BEV) representations provide a structured, top-down abstraction of the vehicle's surroundings and have become a key input modality for Behavioral Cloning (BC) policies. While ground-truth BEV maps are readily available in simulation, real-world deployment requires replacing them with camera-predicted counterparts - a substitution that introduces perceptual errors whose downstream impact on closed-loop driving performance is not well understood. In this work, we investigate the use of Cross-View Transformer (CVT)-predicted BEV maps as direct policy inputs for a BC agent in the CARLA simulator. We propose a six-channel BEV representation covering road surface, planned route, lane boundaries, vehicles, pedestrians, and traffic lights, and introduce a Kernel Density Estimation (KDE) weighting scheme that rebalances the segmentation loss towards underrepresented driving maneuvers such as curves and intersections. Closed-loop evaluation across two CARLA towns shows that the KDE-weighted model is the only predicted-BEV agent to complete a full episode without infractions, despite not achieving the highest aggregate IoU. This discrepancy reveals that global segmentation metrics are poor proxies for driving performance: what determines navigation success is prediction quality at geometrically critical locations, and the route channel emerges as the primary bottleneck for reliable agent navigation under predicted BEV inputs.

0 citationsRead paper
Recent publications

Latest Papers

Closed-Loop Evaluation of Bird's-Eye-View Maps from Cross-View Transformers as Inputs to Behavior-Cloning Policies

Sep 05, 2026

In autonomous driving, Bird's-Eye View (BEV) representations provide a structured, top-down abstraction of the vehicle's surroundings and have become a key input modality for Behavioral Cloning (BC) policies. While ground-truth BEV maps are readily available in simulation, real-world deployment requires replacing them with camera-predicted counterparts - a substitution that introduces perceptual errors whose downstream impact on closed-loop driving performance is not well understood. In this work, we investigate the use of Cross-View Transformer (CVT)-predicted BEV maps as direct policy inputs for a BC agent in the CARLA simulator. We propose a six-channel BEV representation covering road surface, planned route, lane boundaries, vehicles, pedestrians, and traffic lights, and introduce a Kernel Density Estimation (KDE) weighting scheme that rebalances the segmentation loss towards underrepresented driving maneuvers such as curves and intersections. Closed-loop evaluation across two CARLA towns shows that the KDE-weighted model is the only predicted-BEV agent to complete a full episode without infractions, despite not achieving the highest aggregate IoU. This discrepancy reveals that global segmentation metrics are poor proxies for driving performance: what determines navigation success is prediction quality at geometrically critical locations, and the route channel emerges as the primary bottleneck for reliable agent navigation under predicted BEV inputs.

0 citationsRead paper