A Bayesian Framework for Active Tactile Object Recognition, Pose Estimation and Shape Transfer Learning

As humans can explore and understand the world through active touch, similar capability is desired for robots. In this paper, we address the problem of active tactile object recognition, pose estimation and shape transfer learning, where a customized particle filter (PF) and Gaussian process implicit surface (GPIS) is combined in a unified Bayesian framework. Upon new tactile input, the customized PF updates the joint distribution of the object class and object pose while tracking the novelty of the object. Once a novel object is identified, its shape will be reconstructed using GPIS. By grounding the prior of the GPIS with the maximum-a-posteriori (MAP) estimation from the PF, the knowledge about known shapes can be transferred to learn novel shapes. An exploration procedure based on global shape estimation is proposed to guide active data acquisition and terminate the exploration upon sufficient information. Through experiments in simulation, the proposed framework demonstrated its effectiveness and efficiency in estimating object class and pose for known objects and learning novel shapes. Furthermore, it can recognize previously learned shapes reliably.

💡 Research Summary

This paper presents a unified Bayesian framework that simultaneously tackles tactile object recognition, 6‑DOF pose estimation, and shape learning for novel objects. The core of the framework is a customized particle filter (PF) that maintains a joint posterior over object class and pose, and a Gaussian Process Implicit Surface (GPIS) that reconstructs object geometry with uncertainty.

Problem Setting
Robotic tactile sensing provides only local, sparse measurements, making it impossible to determine an object’s identity, pose, and shape from a single touch. Consequently, active exploration is required, and the robot must also decide whether the object belongs to a known set or is a new instance that needs to be learned. Existing works typically address these tasks separately, either assuming a known object model for pose estimation or using fixed priors for shape reconstruction without novelty detection.

Bayesian Formulation
The latent variable (z =