Search

Bringing tactile sensation to robotic hands will allow for more effective grasping, along with the wide range of benefits of human-like touch. Here we present a 3D-printed, three-fingered tactile robot hand comprising an OpenHand Model O customized to house a TacTip soft biomimetic tactile sensor in the distal phalanx of each finger. We expect that combining the grasping capabilities of this underactuated hand with sophisticated tactile sensing will result in an effective platform for robot hand research -- the Tactile Model O (T-MO). The design uses three JeVois machine vision systems, each comprising a miniature camera in the tactile fingertip with a processing module in the base of the hand. To evaluate the capabilities of the T-MO, we benchmark its grasping performance using the Gripper Assessment Benchmark on the YCB object set. Tactile sensing capabilities are evaluated by performing tactile object classification on 26 objects and predicting whether a grasp will successfully lift each object. Results are consistent with the state of the art, taking advantage of advances in deep learning applied to tactile image outputs. Overall, this work demonstrates that the T-MO is an effective platform for robot hand research and we expect it to open-up a range of applications in autonomous object handling. Supplemental video: https://youtu.be/RTcCpgffCrQ., 15 pages, 10 figures, 3 tables

Hyperpolarized (129)Xe MRI has emerged as a novel means to evaluate pulmonary function via 3D mapping of ventilation, interstitial barrier uptake, and RBC transfer. However, the physiological interpretation of these measurements has yet to be firmly established. Here, we propose a model that uses the three components of (129)Xe gas-exchange MRI to estimate accessible alveolar volume (V(A)), membrane conductance, and capillary blood volume contributions to DL(CO). (129)Xe ventilated volume (VV) was related to V(A) by a scaling factor k(V) = 1.47 with 95% confidence interval [1.42, 1.52], relative (129)Xe barrier uptake (normalized by the healthy reference value) was used to estimate the membrane-specific conductance coefficient k(B) = 10.6 [8.6, 13.6] mL/min/mmHg/L, whereas normalized RBC transfer was used to calculate the capillary blood volume-specific conductance coefficient k(R) = 13.6 [11.4, 16.7] mL/min/mmHg/L. In this way, the barrier and RBC transfer per unit volume determined the transfer coefficient K(CO), which was then multiplied by image-estimated V(A) to obtain DL(CO). The model was built on a cohort of 41 healthy subjects and 101 patients with pulmonary disorders. The resulting (129)Xe-derived DL(CO) correlated strongly (R(2) = 0.75, P < 0.001) with the measured values, a finding that was preserved within each individual disease cohort. The ability to use (129)Xe MRI measures of ventilation, barrier uptake, and RBC transfer to estimate each of the underlying constituents of DL(CO) clarifies the interpretation of these images while enabling their use to monitor these aspects of gas exchange independently and regionally. NEW & NOTEWORTHY The diffusing capacity for carbon monoxide (DL(CO)) is perhaps one of the most comprehensive physiological measures used in pulmonary medicine. Here, we spatially resolve and estimate its key components—accessible alveolar volume, membrane, and capillary blood volume conductances—using hyperpolarized (129)Xe MRI of ventilation, interstitial barrier uptake, and red blood cell transfer. This image-derived DL(CO) correlates strongly with measured values in 142 subjects with a broad range of pulmonary disorders.

Artificial touch would seem well-suited for Reinforcement Learning (RL), since both paradigms rely on interaction with an environment. Here we propose a new environment and set of tasks to encourage development of tactile reinforcement learning: learning to type on a braille keyboard. Four tasks are proposed, progressing in difficulty from arrow to alphabet keys and from discrete to continuous actions. A simulated counterpart is also constructed by sampling tactile data from the physical environment. Using state-of-the-art deep RL algorithms, we show that all of these tasks can be successfully learnt in simulation, and 3 out of 4 tasks can be learned on the real robot. A lack of sample efficiency currently makes the continuous alphabet task impractical on the robot. To the best of our knowledge, this work presents the first demonstration of successfully training deep RL agents in the real world using observations that exclusively consist of tactile images. To aid future research utilising this environment, the code for this project has been released along with designs of the braille keycaps for 3D printing and a guide for recreating the experiments. A brief video summary is also available at https://youtu.be/eNylCA2uE_E., Comment: Accepted in RAL and IROS 2020