Research Analyzer
← Back ICRA 2026

A Bayesian Modeling Framework for Estimation and Ground Segmentation of Cluttered Staircases

Prasanna Sriganesh, Burhanuddin Shirose, Matthew Travers

PDF

AI summary

Key figure (auto-extracted from paper)
A Bayesian framework with a novel split state-space model enables robots to robustly estimate and segment cluttered staircases despite occlusions and sensor noise.
Staircase estimation Bayesian inference Robot navigation Ground segmentation State-space modeling Legged robots

Problem

Autonomous robots struggle to safely navigate cluttered or occluded staircases due to limited field-of-view, sensor noise, and missing data, which can cause misinterpretation of obstacles as steps.

Approach

The method fuses noisy sensor measurements with a predictive split state-space model using an Extended Kalman Filter to continuously estimate staircase geometry and segment safe ground regions.

Key results

  • Novel split state-space model for large-scale staircase representation
  • Robust MAP estimation pipeline using Bayesian inference and EKF
  • Stair surface segmentation algorithm for cluttered environments
  • 67-89% reduction in parameter errors and 30% reduction in location error vs. baselines

Why it matters

Enables safer and more reliable autonomous navigation for legged robots in complex, real-world environments with occluded or cluttered stairs.

Abstract

Autonomous robot navigation in complex environ- ments requires robust perception as well as high-level scene understanding due to perceptual challenges, such as occlusions, and uncertainty introduced by robot movement. For example, a robot climbing a cluttered staircase can misinterpret clutter as a step, misrepresenting the state and compromising safety. This requires robust state estimation methods capable of inferring the underlying structure of the environment even from incomplete sensor data. In this paper, we introduce a novel method for robust state estimation of staircases. To address the challenge of perceiv- ing occluded staircases extending beyond the robot’s field-of-view, our approach combines an infinite-width staircase representation with a finite endpoint state to capture the overall staircase structure. This representation is integrated into a Bayesian infer- ence framework to fuse noisy measurements enabling accurate estimation of staircase location even with partial observations and occlusions. Additionally, we present a segmentation algorithm that works in conjunction with the staircase estimation pipeline to accurately identify clutter-free regions on a staircase. Our method is extensively evaluated on real robots across diverse staircases, demonstrating significant improvements in estimation accuracy and segmentation performance compared to baseline approaches.

Index terms

Object Detection Segmentation and Categorization Probabilistic Inference Field Robots

Related papers