Search papers, labs, and topics across Lattice.
This paper analyzes the distributional stability of tangent-linearized Gaussian inference on smooth manifolds, relevant to robotics applications. It derives non-asymptotic Wasserstein-2 ($W_2$) stability bounds for projection marginalization and surface-measure conditioning, separating local geometric distortion from nonlocal tail leakage. Experiments on circle and planar-pushing tasks validate predicted calibration transitions and identify normal-direction uncertainty as the primary failure mode when locality assumptions break down.
Tangent-linearized Gaussian inference on manifolds breaks down predictably, with a simple diagnostic that tells you when to switch to more complex methods.
Gaussian inference on smooth manifolds is central to robotics, but exact marginalization and conditioning are generally non-Gaussian and geometry-dependent. We study tangent-linearized Gaussian inference and derive explicit non-asymptotic $W_2$ stability bounds for projection marginalization and surface-measure conditioning. The bounds separate local second-order geometric distortion from nonlocal tail leakage and, for Gaussian inputs, yield closed-form diagnostics from $(μ,Σ)$ and curvature/reach surrogates. Circle and planar-pushing experiments validate the predicted calibration transition near $\sqrt{\|Σ\|_{\mathrm{op}}}/R\approx 1/6$ and indicate that normal-direction uncertainty is the dominant failure mode when locality breaks. These diagnostics provide practical triggers for switching from single-chart linearization to multi-chart or sample-based manifold inference.
