Abstract:This paper proposes a novel method for estimating the set of plausible poses of a rigid object from a set of points with volumetric information, such as whether each point is in free space or on the surface of the object. In particular, we study how pose can be estimated from force and tactile data arising from contact. Using data derived from contact is challenging because it is inherently less information-dense than visual data, and thus the pose estimation problem is severely under-constrained when there are few contacts. Rather than attempting to estimate the true pose of the object, which is not tractable without a large number of contacts, we seek to estimate a plausible set of poses which obey the constraints imposed by the sensor data. Existing methods struggle to estimate this set because they are either designed for single pose estimates or require informative priors to be effective. Our approach to this problem, Constrained pose Hypothesis Set Elimination (CHSEL), has three key attributes: 1) It considers volumetric information, which allows us to account for known free space; 2) It uses a novel differentiable volumetric cost function to take advantage of powerful gradient-based optimization tools; and 3) It uses methods from the Quality Diversity (QD) optimization literature to produce a diverse set of high-quality poses. To our knowledge, QD methods have not been used previously for pose registration. We also show how to update our plausible pose estimates online as more data is gathered by the robot. Our experiments suggest that CHSEL shows large performance improvements over several baseline methods for both simulated and real-world data.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to estimate a set of possible poses of a rigid object given a set of points containing volume information (for example, whether these points are in free space or on the object surface). Specifically, the author has studied how to estimate poses from forces and tactile data generated by contact. Since contact data is inherently less information - dense than visual data, the pose - estimation problem is severely under - constrained when there are few contact points. Traditional methods are difficult to estimate this set of possible poses because they are either designed for single - pose estimation or require informative priors to be effective. To solve these problems, the author proposes a method named **Constrained pose Hypothesis Set Elimination (CHSEL)**, which has the following three key properties: 1. **Consider volume information**: Allows us to take into account the known free space. 2. **Use a new differentiable volume cost function**: Utilize powerful gradient - based optimization tools. 3. **Use Quality Diversity (QD) optimization method**: Generate a set of high - quality and diverse pose estimates. Through these methods, CHSEL can estimate a set of possible poses that conform to the sensor - data constraints when the contact data is limited. In addition, the paper also shows how to update pose estimates online as the robot collects more data. Experimental results show that CHSEL significantly outperforms several baseline methods on both simulated and real - world data.

CHSEL: Producing Diverse Plausible Pose Estimates from Contact and Free Space Data

Pose Estimation and Neural Implicit Reconstruction Towards Non-Cooperative Spacecraft Without Offline Prior Information

MoreFusion: Multi-object Reasoning for 6D Pose Estimation from Volumetric Fusion

ContourPose: Monocular 6D Pose Estimation Method for Reflective Texture-Less Metal Parts

Robust And Accurate Multiple-Camera Pose Estimation Toward Robotic Applications

Robust 3D Human Pose Estimation from Single Images or Video Sequences

ContactArt: Learning 3D Interaction Priors for Category-level Articulated Object and Hand Poses Estimation

A Method for Unseen Object Six Degrees of Freedom Pose Estimation Based on Segment Anything Model and Hybrid Distance Optimization

Nothing But Geometric Constraints: A Model-Free Method for Articulated Object Pose Estimation

A Multi-Hypothesis Approach to Pose Ambiguity in Object-Based SLAM

6D Pose Estimation of Industrial Parts Based on Point Cloud Geometric Information Prediction for Robotic Grasping

Robust Estimation of 3D Human Poses from a Single Image

Joint Hand and Object Pose Estimation from a Single RGB Image Using High-level 2D Constraints

CHAMP: Conformalized 3D Human Multi-Hypothesis Pose Estimators

Resolving 3D Human Pose Ambiguities with 3D Scene Constraints

Improving Robustness and Accuracy via Relative Information Encoding in 3D Human Pose Estimation

3D Foundation Models Enable Simultaneous Geometry and Pose Estimation of Grasped Objects

SD-Pose: facilitating space-decoupled human pose estimation via adaptive pose perception guidance

Learning a Contact Potential Field for Modeling the Hand-Object Interaction

Embodied Scene-aware Human Pose Estimation

DSC-PoseNet: Learning 6DoF Object Pose Estimation via Dual-scale Consistency