基于点阵模型的惯性融合导航技术

Inertial Fusion Navigation Technology Based on the Lattice Model

摘要: 为解决单惯性导航系统中锚点引起的空间漂移问题，通常采用多传感器系统，通过不同传感器间的相互锚定机制有效抑制空间漂移。然而，传统方法在提升导航精度和降低数据分析复杂度方面效果有限。本文提出了一种基于物理类比的、有前景的探索性方法，该方法类比晶体物理的点阵结构，将多个惯性传感器抽象为点阵中的格点，利用格点间的相对约束关系构建对称模型，通过多传感器间的相对位置约束实现漂移误差的自校正，进而抑制导航过程中的空间漂移。本文分别设计多组模拟实验与一组真实场景实验对所提方法进行验证，实验结果表明，相较于传统单惯性导航方法与常规多传感器融合方法，该方法可以在控制算法复杂度的前提下，有效提升定位导航的精度，为多惯性传感器协同导航提供了一种新的可行思路。该方法为惯性导航系统的多传感器融合提供了新的视角与方法论，有望推动该领域进一步的技术进步。

Abstract: To address spatial drift caused by anchor points in single-inertial navigation systems, multi-sensor systems are typically employed, effectively suppressing spatial drift through mutual anchoring mechanisms among sensors. However, traditional approaches demonstrate limited effectiveness in enhancing navigation accuracy and reducing data analysis complexity. This paper proposes a promising exploratory method based on physical analogy, which mimics the lattice structure of crystal physics by abstracting multiple inertial sensors as grid points within a lattice. By leveraging the relative constraint relationships between grid points to construct a symmetric model, the method achieves self-correction of drift errors through positional constraints among sensors, thereby mitigating spatial drift during navigation. The proposed method is validated through multiple simulated experiments and one real-world scenario experiment. Results demonstrate that, compared to conventional single-inertial navigation methods and standard multi-sensor fusion approaches, this method significantly improves positioning and navigation accuracy while maintaining comparable control algorithm complexity, offering a novel and feasible strategy for collaborative multi-inertial sensor navigation. This approach provides fresh perspectives and methodologies for multi-sensor fusion in inertial navigation systems, poised to drive further technological advancements in the field.