PCARNN-DCBF: Minimal-Intervention Geofence Enforcement for Ground Vehicles

TL;DR

This paper introduces PCARNN-DCBF, a novel control pipeline combining physics-encoded neural networks with a discrete control barrier function, enabling reliable, real-time geofence enforcement for ground vehicles with high fidelity and structural guarantees.

Contribution

The paper presents a new structure-preserving neural network combined with a preview-based control barrier function for verifiable, high-fidelity geofence enforcement in ground vehicles.

Findings

Outperforms analytical and unstructured neural baselines in CARLA simulations.

Ensures real-time enforcement of polygonal keep-in constraints.

Maintains control-affine structure for reliable optimization.

Abstract

Runtime geofencing for ground vehicles is rapidly emerging as a critical technology for enforcing Operational Design Domains (ODDs). However, existing solutions struggle to reconcile high-fidelity learning with the structural requirements of verifiable control. We address this by introducing PCARNN-DCBF, a novel pipeline integrating a Physics-encoded Control-Affine Residual Neural Network with a preview-based Discrete Control Barrier Function. Unlike generic learned models, PCARNN explicitly preserves the control-affine structure of vehicle dynamics, ensuring the linearity required for reliable optimization. This enables the DCBF to enforce polygonal keep-in constraints via a real-time Quadratic Program (QP) that handles high relative degree and mitigates actuator saturation. Experiments in CARLA across electric and combustion platforms demonstrate that this structure-preserving approach significantly outperforms analytical and unstructured neural baselines.