MapEx: Indoor Structure Exploration with Probabilistic Information Gain from Global Map Predictions

2024-09-23

arXiv.org

AI Summary

Existing autonomous exploration methods for structured indoor environments suffer from three key limitations: failure to exploit environmental regularity, overreliance on high-fidelity predicted maps, and neglect of sensor visibility constraints. To address these, this paper proposes a probabilistic information gain evaluation framework that jointly models observable regions and uncertainty in predicted maps. It is the first approach to unify multi-sample predictive map variance with geometric visibility constraints, thereby overcoming restrictive assumptions inherent in conventional frontier-based and deterministic prediction-based methods. The framework integrates deep learning–based map prediction, Monte Carlo sampling for uncertainty distribution estimation, a probabilistic sensor model, and exact visibility computation. Evaluated on the KTH real-world dataset, our method achieves a 12.4% improvement over representative map-prediction–guided exploration approaches and a 25.4% gain over the state-of-the-art frontier-based method, demonstrating significant advances in both exploration efficiency and robustness.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Intelligent Robots: Localization, Mapping, and Navigation
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Exploration is a critical challenge in robotics, centered on understanding unknown environments. In this work, we focus on robots exploring structured indoor environments which are often predictable and composed of repeating patterns. Most existing approaches, such as conventional frontier approaches, have difficulty leveraging the predictability and explore with simple heuristics such as `closest first'. Recent works use deep learning techniques to predict unknown regions of the map, using these predictions for information gain calculation. However, these approaches are often sensitive to the predicted map quality or do not reason over sensor coverage. To overcome these issues, our key insight is to jointly reason over what the robot can observe and its uncertainty to calculate probabilistic information gain. We introduce MapEx, a new exploration framework that uses predicted maps to form probabilistic sensor model for information gain estimation. MapEx generates multiple predicted maps based on observed information, and takes into consideration both the computed variances of predicted maps and estimated visible area to estimate the information gain of a given viewpoint. Experiments on the real-world KTH dataset showed on average 12.4% improvement than representative map-prediction based exploration and 25.4% improvement than nearest frontier approach.

Problem

• Explores structured indoor environments using probabilistic information gain
• Overcomes sensitivity to predicted map quality in existing approaches
• Improves exploration efficiency with predicted maps and sensor coverage

Innovation

• Uses predicted maps for probabilistic sensor modeling
• Generates multiple maps considering observed variances
• Estimates information gain via visibility and uncertainty