Autonomous Exploration for Shape Reconstruction and Measurement Via Informative Contact-Guided Planning
Feiyu Zhao, Chenxi Xiao
AI summary
Problem
Traditional coordinate measuring machines rely on CAD model priors and predefined probing paths, limiting their use for inspecting freeform, damaged, or legacy parts. They also lack the autonomy and dense data acquisition needed for accurate shape reconstruction in unknown environments.
Approach
The system incrementally probes unknown objects using tactile contact, guided by a dual Gaussian Process architecture that simultaneously estimates surface shape and exploration uncertainty. A hybrid motion planner adaptively switches between local sliding, global redirection, and contact recovery to efficiently sample informative regions.
Key results
- CAD-free tactile framework enabling dense 3D shape reconstruction from contact data alone
- Dual Gaussian Process model (ER-GPIS) decoupling shape reconstruction and uncertainty estimation for robust sparse-data handling
- Hybrid motion planner integrating local sliding, global exploration, and contact recovery policies
- Real-world validation showing accurate reconstruction of diverse geometries, including internal cavities, outperforming tactile baselines
Why it matters
Enables precise, autonomous geometric inspection and reverse engineering of industrial parts where CAD models and visual sensors are unavailable or ineffective.
Abstract
Coordinate Measuring Machines (CMMs) are widely used for high-precision inspection of industrial parts, particularly in scenarios where visual systems are ineffective or cost-prohibitive. However, conventional CMMs rely on CAD model priors and user-defined probing paths, which limit their applicability and efficiency in measuring freeform parts. To overcome these limitations, we present a fully autonomous, CAD model-free, tactile-based framework that enables dense 3D shape reconstruction to facilitate subsequent measurements. Our approach leverages a dual Gaussian Process Implicit Sur- face architecture, termed Exploration-Reconstruction GPIS (ER- GPIS), which enables both high-fidelity shape reconstruction and uncertainty estimation on the object’s surface. A hybrid exploration motion planner is then employed to adaptively sample surface geometries by integrating local surface exploration, global exploration, and contact recovery policies for robust shape estimation. Extensive real-world experiments demonstrate that the proposed method effectively reconstructs object geometries across diverse shapes, highlighting its ability to autonomously reconstruct and measure both surfaces and internal features without relying on CAD model priors. Our project webpage is available at https://aesrm.github.io/