Abstract
Modeling deformable contact is a well-known problem in soft robotics and is particularly challenging for compliant interfaces that permit large deformations. We present a model for the behavior of a highly deformable dense geometry sensor in its interaction with objects; the forward model predicts the elastic deformation of a mesh given the pose and geometry of a contacting rigid object. We use this model to develop a fast approximation to solve the inverse problem: estimating the contact patch when the sensor is deformed by arbitrary objects. This inverse model can be easily identified through experiments and is formulated as a sparse Quadratic Program (QP) that can be solved efficiently online. The proposed model serves as the first stage of a pose estimation pipeline for robot manipulation. We demonstrate the proposed inverse model through real-time estimation of contact patches on a contact-rich manipulation problem in which oversized fingers screw a nut onto a bolt, and as part of a complete pipeline for pose-estimation and tracking based on the Iterative Closest Point (ICP) algorithm. Our results demonstrate a path towards realizing soft robots with highly compliant surfaces that perform complex real-world manipulation tasks.
Highlights
W E IMAGINE a future with robots that are able to make contact anywhere on their bodies in order to successfully execute tasks
We present two key contributions towards addressing these challenges: (i) We present a forward model based on first principles of continuum mechanics to describe a highlydeformable air-filled membrane that makes contact with a rigid object of a given geometry. (ii) We utilize this model to solve the inverse problem of identifying the contact patch based solely on the depth information from the sensor - we develop an approximate formulation to solve this problem using a sparse convex Quadratic Program (QP) which renders it solvable in real-time
In this letter we have presented a model for the behavior of highly-deformable dense-geometry sensors
Summary
W E IMAGINE a future with robots that are able to make contact anywhere on their bodies in order to successfully execute tasks. As a visuo-tactile sensor, it provides a high resolution depth image of the deformed contact surface In previous work, this sensor has been used for object classification and pose estimation [6]. We present two key contributions towards addressing these challenges: (i) We present a forward model based on first principles of continuum mechanics to describe a highlydeformable air-filled membrane that makes contact with a rigid object of a given geometry. Its ability to deform around a contacting object more deeply than thinner gel-based sensors results in a larger portion of the object’s geometry being captured in the output depth map This sensor’s high resolution depth sensing, along with its ability to provide a large, high-friction contact patch make it an ideal contacting surface for manipulators. We show in subsequent sections how this formulation can be employed within both a forward simulation as well as an inverse model that estimates the contact patch between the soft sensor and the external world
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have