Mitigating Error Accumulation in Continuous Navigation via Memory-Augmented Kalman Filtering

2026-01-30

arXiv.org

AI Summary

This work addresses the severe degradation in trajectory prediction accuracy caused by error accumulation and state drift inherent in recursive position updates during continuous navigation. The authors formulate navigation as a recursive Bayesian state estimation problem and propose NeuroKalman, a framework that decouples the process into two stages: prior prediction based on motion dynamics and likelihood correction informed by historical observations. A memory mechanism is introduced to dynamically refine state representations. Innovatively, the method establishes a mathematical linkage between kernel density estimation and attention-based retrieval, enabling history-guided correction via anchor points without requiring gradient updates, thereby effectively mitigating long-term drift. Evaluated on the TravelUAV benchmark, NeuroKalman achieves significant performance gains over strong baselines with only 10% of the training data for fine-tuning, demonstrating its efficacy in controlling error accumulation.

Abstract

Continuous navigation in complex environments is critical for Unmanned Aerial Vehicle (UAV). However, the existing Vision-Language Navigation (VLN) models follow the dead-reckoning, which iteratively updates its position for the next waypoint prediction, and subsequently construct the complete trajectory. Then, such stepwise manner will inevitably lead to accumulated errors of position over time, resulting in misalignment between internal belief and objective coordinates, which is known as"state drift"and ultimately compromises the full trajectory prediction. Drawing inspiration from classical control theory, we propose to correct for errors by formulating such sequential prediction as a recursive Bayesian state estimation problem. In this paper, we design NeuroKalman, a novel framework that decouples navigation into two complementary processes: a Prior Prediction, based on motion dynamics and a Likelihood Correction, from historical observation. We first mathematically associate Kernel Density Estimation of the measurement likelihood with the attention-based retrieval mechanism, which then allows the system to rectify the latent representation using retrieved historical anchors without gradient updates. Comprehensive experiments on TravelUAV benchmark demonstrate that, with only 10% of the training data fine-tuning, our method clearly outperforms strong baselines and regulates drift accumulation.

Problem

• error accumulation
• state drift
• continuous navigation
• Vision-Language Navigation
• trajectory prediction

Innovation

• Memory-Augmented Kalman Filtering
• Error Accumulation Mitigation
• Recursive Bayesian Estimation
• Attention-based Retrieval
• State Drift Correction