Conflict-Aware Active Perception and Control in 3D Gaussian Splatting Fields via Control Barrier Functions

2026-05-19

AI Summary

This work addresses the challenge of active robotic perception in uncertain environments, where maximizing information gain often conflicts with ensuring safe navigation due to high collision risks at informative viewpoints. To reconcile this trade-off, the authors propose a conflict-aware active perception and control framework that leverages 3D Gaussian Splatting for scene representation. Safety is rigorously enforced through a control barrier function incorporating Average Value-at-Risk (AV@R) as a risk measure, while a novel risk-aware expected information gain metric guides viewpoint selection. These components are unified within a safety-critical quadratic program augmented with slack variables via a perception barrier function. Simulations demonstrate that the proposed method significantly improves information acquisition efficiency while strictly guaranteeing safety, outperforming existing 3DGS-based active perception approaches.

Abstract

Active perception in uncertain environments requires robots to navigate safely while acquiring informative observations to reduce map uncertainty. These objectives inherently conflict, as informative viewpoints often lie near uncertain regions with higher collision risk. To address this challenge, we develop a conflict-aware active perception and control framework for robotic systems operating in environments represented by 3D Gaussian Splatting (3DGS). Safety is enforced using a Control Barrier Function (CBF) derived from an Average Value-at-Risk AV@R collision-risk metric that accounts for geometric uncertainty and guarantees forward invariance of a safe set. To improve perception, we propose a risk-aware Expected Information Gain (EIG) formulation for selecting the next-best-view and introduce perception barrier functions that align the camera orientation with the local information-ascent direction. To obtain a tractable formulation for these conflicting safety and perception objectives, we propose a unified safety-critical, perception-aware quadratic program that enforces safety as a hard constraint while relaxing perception constraints through slack variables. Simulation results demonstrate that the proposed method improves both safety and information acquisition compared to existing 3DGS-based approaches.

Problem

• active perception
• safety-perception conflict
• 3D Gaussian Splatting
• collision risk
• uncertain environments

Innovation

• Control Barrier Functions
• 3D Gaussian Splatting
• Active Perception
• Risk-Aware Planning
• Expected Information Gain