Inverse Covariance Intersection (ICI) Overview

• Inverse Covariance Intersection (ICI) is a data fusion method that combines incompatible Gaussian estimates using convex optimization of fusion weights.
• It optimally selects weights by minimizing criteria like trace or log-determinant to balance conservativeness and informativeness, ensuring rigorous uncertainty bounds.
• ICI extends to distributed filtering on Lie groups and partial-knowledge models, enabling robust state estimation in sensor networks and target tracking applications.

Inverse Covariance Intersection (ICI) is a data fusion methodology for combining incompatible or correlatively ambiguous Gaussian estimates, ensuring estimation consistency and providing less conservatism than standard Covariance Intersection (CI). ICI is central to distributed filtering and estimation, particularly in sensor networks for both vector spaces and matrix Lie groups, as well as in advanced settings such as partial knowledge regimes (e.g., common information or common noise scenarios). The method enables robust fusion of information under unknown or only partially known cross-correlation, extending standard fusion techniques to contexts where rigorous consistency guarantees are required (Ruan et al., 2024, Ajgl et al., 6 Jun 2025).

1. Consistent Fusion under Unknown Correlations

ICI addresses the core challenge of combining two or more Gaussian estimates $x_1 \sim \mathcal{N}(x_1, P_1)$ and $x_2 \sim \mathcal{N}(x_2, P_2)$ whose cross-covariances are unknown, which precludes direct optimal linear fusion. Naive consensus approaches may yield fused covariances that underestimate the true post-fusion uncertainty, violating consistency. ICI constructs a fused Gaussian $\mathcal{N}(x_f, P_f)$ such that $P_f$ upper-bounds the actual error covariance for any admissible correlation, and aims to be less conservative than standard CI while remaining architecturally straightforward for distributed deployment.

ICI constructs the fused covariance as

$$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$

where $\omega \in [0,1]$ is a fusion weight satisfying $\omega \geq 0$, $1-\omega \geq 0$, and $\omega + (1-\omega) = 1$. The corresponding fused mean is

$$x_f = P_f [\omega P_1^{-1} x_1 + (1-\omega) P_2^{-1} x_2].$$

Weight selection relies on convex scalarizations, typically minimizing either $x_2 \sim \mathcal{N}(x_2, P_2)$0 or $x_2 \sim \mathcal{N}(x_2, P_2)$1, resulting in a convex one-dimensional program in $x_2 \sim \mathcal{N}(x_2, P_2)$2 (Ruan et al., 2024).

2. Weight Selection and Optimization Criteria

The effectiveness of ICI in balancing conservativeness and informativeness hinges on optimal selection of $x_2 \sim \mathcal{N}(x_2, P_2)$3. Canonical criteria are:

• Minimize trace: $x_2 \sim \mathcal{N}(x_2, P_2)$4
• Minimize log-determinant: $x_2 \sim \mathcal{N}(x_2, P_2)$5

Both criteria yield a convex optimization problem that can be solved efficiently. These scalarizations enable implementers to tune fusion conservativeness or informativeness to application demands while preserving rigorous error bounds.

3. Extension to Matrix Lie Groups

In distributed estimation over matrix Lie groups (e.g., $x_2 \sim \mathcal{N}(x_2, P_2)$6), ICI generalizes to handle manifold-valued state spaces and associated local uncertainties defined on Lie algebras. Each agent $x_2 \sim \mathcal{N}(x_2, P_2)$7 stores an estimate $x_2 \sim \mathcal{N}(x_2, P_2)$8 and associated tangent-space covariance $x_2 \sim \mathcal{N}(x_2, P_2)$9. The right-invariant local error is represented as $\mathcal{N}(x_f, P_f)$0 via $\mathcal{N}(x_f, P_f)$1, with $\mathcal{N}(x_f, P_f)$2.

For a neighborhood $\mathcal{N}(x_f, P_f)$3, agents construct error-difference vectors

$\mathcal{N}(x_f, P_f)$4

which, under Baker–Campbell–Hausdorff (BCH) linearization, satisfy $\mathcal{N}(x_f, P_f)$5, serving as noisy observations of $\mathcal{N}(x_f, P_f)$6 with noise covariance $\mathcal{N}(x_f, P_f)$7. Multi-fusion ICI operates in the Lie algebra:

$\mathcal{N}(x_f, P_f)$8

This Lie group ICI extension enables consistent uncertainty fusion in a broad class of geometric estimation problems and supports integration with invariant filtering architectures (Ruan et al., 2024).

4. Integration with Distributed Filtering and the Invariant UKF

ICI is naturally incorporated in distributed sensor network filters. In the diffusion-based distributed invariant Unscented Kalman Filter (DIUKF-ICI), each agent performs the following per timestep:

• A: Local UKF propagation and measurement update yielding $\mathcal{N}(x_f, P_f)$9.
• B: Incremental update incorporating neighbor measurements $P_f$0.
• C: Diffusion-fusion via extended ICI: exchange $P_f$1; compute $P_f$2; solve convex program for $P_f$3; compute $P_f$4 via ICI formulas; update $P_f$5, $P_f$6.

This structure is robust to intermittent measurements and time-varying communication topologies. Agents omit non-participating neighbors or assign vanishing weights to estimates with infinite (or very large) covariances. All fusion and weight constraints are localized to the current neighborhood, ensuring consistency across dynamic networks (Ruan et al., 2024).

5. Dual ICI and Partial Knowledge Models

Beyond total ignorance of inter-estimate correlation, ICI admits dual formulations that exploit partial prior knowledge, notably the common-noise regime. For two unbiased estimates $P_f$7 of $P_f$8 with marginal covariances $P_f$9, $$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$0 and a joint covariance $$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$1 where $$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$2, dual-ICI constructs an upper bound $$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$3 and applies fusion with weights enforcing $$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$4.

The upper bound is parameterized as

$$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$5

where $$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$6. The fused covariance bound is

$$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$7

with $$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$8. For the optimal $$P_f = [\omega P_1^{-1} + (1-\omega) P_2^{-1}]^{-1},$$9, $\omega \in [0,1]$0 coincides with the standard CI bound.

Significantly, although the union of dual-ICI bounds across admissible $\omega \in [0,1]$1 strictly improves the union of CI bounds with suboptimal weights, no single dual-ICI bound outperforms the CI bound at its optimal $\omega \in [0,1]$2 (Ajgl et al., 6 Jun 2025). A plausible implication is that partial knowledge models can be exploited to tighten covariance bounds in multi-criteria or adaptive settings, but the gain is not realized in standard scalarized optimal fusion.

6. Properties, Guarantees, and Practical Considerations

ICI ensures the fused covariance is always a valid upper bound: $\omega \in [0,1]$3. For the distributed Lie group setting, Theorem 1 of (Ruan et al., 2024) establishes that, under uniform bounds on local covariances and measurement noise, the sequences of local and fused covariances remain uniformly bounded, guaranteeing mean-square bounded error trajectories.

Algorithmic robustness to intermittent observations is maintained by omitting absent neighbors or assigning vanishing fusion weights, while time-varying network topologies are handled via dynamic neighborhoods $\omega \in [0,1]$4 and properly adjusted weight constraints. The convexity of the weight selection criteria simplifies distributed implementations.

7. Applications and Limitations

ICI underlies robust distributed state estimation in sensor networks, particularly where states live on Lie groups (e.g., target tracking in 3D environments). Its deployment in the DIUKF-ICI achieves bounded estimation error under extensive Monte-Carlo testing, even with communication failures and sporadic sensing (Ruan et al., 2024).

In the context of partial correlation knowledge, dual-ICI offers a family of tighter upper bounds for regimes such as common-noise, though improvements are realized only in the ensemble of possible bounds or for suboptimal fusion weights, not at the single optimal scalarized bound (Ajgl et al., 6 Jun 2025). Extending dual-ICI to more agents, partial state measurements, and dynamic fusion remains an active research direction.