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 robots 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 (AVaR) 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 a perception barrier that aligns the camera orientation with the local information-ascent direction.
To obtain a tractable formulation of conflicting safety and perception objectives, we propose a unified safety-critical quadratic program, where safety is enforced as a hard constraint and perception is incorporated via a slack relaxation. Simulation results demonstrate that the proposed method improves both safety and information acquisition compared to existing 3DGS-based approaches.
@inproceedings{khass2026conflict,
title = {Conflict-Aware Active Perception and Control in 3D Gaussian Splatting Fields via Control Barrier Functions},
author = {Khass, Amirhossein Mollaei and Cosse, Athanasios and Pandey, Vivek and Motee, Nader},
booktitle = {Proceedings of the IEEE Conference on Decision and Control (CDC)},
year = {2026},
note = {Submitted}
}