perform probabilistic localization

Real-world sensor data in autonomous driving and healthcare monitoring often exhibit unknown or partially known noise statistics, alongside outliers, biases, drifts, and missing observations—challenging conventional Bayesian filtering. Method: This paper proposes a novel robust nonlinear Bayesian filtering framework that unifies variational inference and particle filtering within a robust Bayesian paradigm. It introduces an anomaly-aware smoothing extension mechanism and establishes a computable Bayesian Cramér–Rao bound (BCRB) for theoretical performance analysis. Furthermore, it extends robustness beyond traditional filtering to learning and generative modeling—e.g., robust diffusion models. Contribution/Results: Evaluated on target tracking, indoor localization, 3D point-cloud registration, and pose-graph optimization, the framework significantly improves estimation accuracy and anomaly tolerance. It provides both theoretical foundations and practical tools for high-reliability intelligent decision-making under complex, real-world sensing uncertainties.

To address erroneous stair structure identification in cluttered staircases caused by occlusions, sensor noise, and limited field-of-view, this paper proposes a robust stair-state estimation and clutter-free region segmentation method. The approach introduces a novel geometric representation of stairs with infinite width coupled with finite endpoint state modeling; designs a Bayesian fusion framework enabling structural inference under partial observations; and jointly optimizes stair-state estimation and clutter-free segmentation. It integrates multi-sensor data with a model-based ground segmentation algorithm. Evaluated on diverse real-world staircase scenarios using a physical robot platform, the method achieves significantly higher stair localization accuracy than baseline approaches and improves clutter-free region segmentation accuracy by 23.6%, thereby effectively supporting safe climbing decision-making.

This work addresses the challenge of exploration and mapping decisions in active SLAM under partial observability by formulating it as a stochastic control problem with incomplete information. The authors propose a non-standard partially observable Markov decision process (POMDP) framework that jointly integrates motion, perception, and map representation. A key innovation is the introduction of an exploration cost function that explicitly captures the geometric structure of the state space to quantify the value of information-gathering actions. Building upon this formulation, they develop a general stochastic control model and derive an approximately optimal policy with theoretical guarantees by combining stochastic control theory and reinforcement learning algorithms. Numerical experiments in representative environments demonstrate the effectiveness of the approach, successfully learning high-performance exploration strategies.

This work proposes a novel framework that integrates classical particle filtering with learning-based methods to address the high training cost and poor interpretability of end-to-end learning in robotic state estimation. Leveraging the Markov assumption, the approach trains a dynamics model using single-step state transitions and implicitly learns the observation model via denoising score matching, thereby approximating the Bayesian filtering equations step-by-step during inference without requiring end-to-end optimization. A key innovation lies in preserving the modular structure of the filter, which enables flexible incorporation of prior knowledge and external sensor models without retraining. Experiments demonstrate that the method achieves accuracy comparable to well-tuned end-to-end baselines in simulation while significantly reducing training complexity and exhibiting superior generalization and compositional capabilities.

In particle filtering, the “prior boundary phenomenon”—estimation failure when target states exceed the limited support of the prior distribution—severely degrades robustness. Method: This paper proposes Diffusion-Enhanced Particle Filtering (DEPF), a novel framework introducing three core mechanisms: adaptive diffusion-based exploration, entropy-driven weight regularization, and dynamic kernel support expansion—enabling online, controllable adaptation of the prior support set. DEPF integrates diffusion process modeling, information-theoretic entropy constraints, kernel density perturbation, and Bayesian resampling to overcome prior boundary limitations while preserving computational efficiency. Contribution/Results: We provide theoretical convergence guarantees for DEPF. Empirical evaluation demonstrates substantial improvements in estimation success rate and accuracy under high-dimensional and non-convex dynamic scenarios; average estimation error decreases by over 40% compared to state-of-the-art baselines.

This work addresses the challenge in conventional tightly coupled SLAM systems, where time handling, geometric association, estimator design, and map updating are highly interdependent, hindering independent optimization. To overcome this, the paper introduces FUSE, a novel framework that decouples core SLAM state estimation components into standardized interfaces for the first time, defining unified mechanisms for observation input, state propagation, update, and query. This modular design enables flexible component substitution. A LiDAR–IMU system built upon FUSE integrates high-frequency IMU propagation, LiDAR-triggered geometric updates, residual filtering, and degeneracy-aware correction. Evaluated on a 418-meter looped corridor sequence, it achieves an end-to-end trajectory error of 1.626 meters, representing a 7.9% reduction in relative error compared to the best-performing baseline, Faster-LIO.

This work addresses the instability of conventional Gaussian sum filters and the high computational cost of particle filters in nonlinear or non-Gaussian state-space models by proposing an Augmented Gaussian Sum Filter (AGSF). The method introduces latent variables and tunable covariance parameters to construct a unified framework that enables continuous interpolation and adaptive switching between Gaussian approximation and particle filtering behaviors. By innovatively integrating augmented Gaussian approximation, adaptive mechanisms, and sequential Monte Carlo principles, AGSF dynamically adjusts its approximation strategy based on the local degree of nonlinearity. Experimental results demonstrate that the proposed approach achieves both efficiency and robustness in target tracking tasks, effectively avoiding failure modes of traditional methods, while toy experiments validate the efficacy of its adaptive mechanism.

This work addresses the challenges of trajectory drift accumulation and computationally expensive global consistency optimization in large-scale SLAM, as well as limitations of existing discrete grid-based submap stitching methods—such as discontinuous gradients and neglect of occupancy uncertainty—by introducing the first continuous probabilistic submap stitching framework. The method jointly optimizes submap poses and a global occupancy field in an implicit log-odds space, compressing raw observations into informative sufficient statistics via sparse Bayesian inference and incorporating a variance-weighting mechanism to preserve posterior uncertainty. It enables analytical Jacobian computation and directly yields an optimal global map with closed-form mean and variance upon pose convergence. Experiments demonstrate significant improvements over state-of-the-art approaches in both simulated and real large-scale environments, achieving higher pose accuracy, enhanced global consistency, greater map compactness, and better-calibrated uncertainty.