Accepted Works 

Authors:

Daeun Song and Young J. Kim

Abstract:

Robotic drawing requires complicated manipulation tasks involving continuous contact with the canvas. This problem is hard since uncertainties arising from modeling, sensor, and computation directly impact drawing quality. In this paper, we address a compliant manipulation problem in the context of robotic drawing using impedance-based control. The system utilizes a seven-degree-of-freedom manipulator capable of position and impedance control and takes Bezier spline  ́ curves as drawing input. To estimate geometric details of the unknown surface during drawing, the system employs incremental and adaptive sampling on the surface through a combination of position and impedance control. The control algorithm then replicates the drawing on any arbitrary surface by impedance-controlling the manipulator. Our results demonstrate the system’s ability to generate visually-appealing and complex pen art on general surfaces without the need for explicit surface reconstruction or visual feedback.

Authors:

Robin Bendfeld and C. David Remy

Abstract:

Accurately measuring deformation in continuously compliant structural elements is essential in various engineering applications. In this study, we investigate methods for sensor placement and sensor fusion for measuring the deformation of a continuously compliant shank in a robotic leg. This compliant shank is part of a new actuation concept, in which the mechanical structure of a robotic leg itself serves as the series elastic element. We explore optimal sensor placement for estimating the foot position during flight and ground contact forces during stance. Furthermore, we introduce a framework for a Kalman filter to estimate the ground reaction forces at the foot contact point. This filter is based on a linearized model of the dynamics of the leg structure.

Authors:

Sergio Aguilera and Seth Hutchinson

Abstract:

There is a wide range of nonholonomic passive wheeled systems that we want mobile manipulators to control. These cart-like systems have a large range of inertial parameters which can change during or between interactions. Our methodology proposes a simplify dynamic model for these cart-like systems and a compliance base controller for mobile manipulators to estimate its parameters, select contact point location and control the system at hand. We present experimental results using a mobile manipulator pushing a shopping cart.

Authors:

Yilin Cai, Shenli Yuan

Abstract:

This paper presents BACH, a compliant robotic hand with active surfaces for secure grasp and in-hand object manipulation. The hand features three identical fingers with an actuated timing belt wrapped around a Fin Ray based compliant finger backbone. Each finger is mounted on an adaptable pivot joint. The hand’s design allows for dexterous in-hand manipulation with great robustness. BACH’s design was optimized and validated through multiple analyses and tests, demonstrating excellent performance with objects of different shapes and sizes.