Elastic Tactile Sensor Glove for Dexterous Teaching by Demonstration.

We present a thin and elastic tactile sensor glove for teaching dexterous manipulation tasks to robots through human demonstration. The entire glove, including the sensor cells, base layer, and electrical connections, is made from soft and stretchable silicone rubber, adapting to deformations under bending and contact while preserving human dexterity. We develop a glove design with five fingers and a palm sensor, revise material formulations for reduced thickness, faster processing and lower cost, adapt manufacturing processes for reduced layer thickness, and design readout electronics for improved sensitivity and battery operation. We further address integration with a multi-camera system and motion reconstruction, wireless communication, and data processing to obtain multimodal reconstructions of human manipulation skills.

Coordinating human-robot collaboration by EEG-based human intention prediction and vigilance control.

In human-robot collaboration scenarios with shared workspaces, a highly desired performance boost is offset by high requirements for human safety, limiting speed and torque of the robot drives to levels which cannot harm the human body. Especially for complex tasks with flexible human behavior, it becomes vital to maintain safe working distances and coordinate tasks efficiently. An established approach in this regard is reactive servo in response to the current human pose. However, such an approach does not exploit expectations of the human's behavior and can therefore fail to react to fast human motions in time. To adapt the robot's behavior as soon as possible, predicting human intention early becomes a factor which is vital but hard to achieve. Here, we employ a recently developed type of brain-computer interface (BCI) which can detect the focus of the human's overt attention as a predictor for impending action. In contrast to other types of BCI, direct projection of stimuli onto the workspace facilitates a seamless integration in workflows. Moreover, we demonstrate how the signal-to-noise ratio of the brain response can be used to adjust the velocity of the robot movements to the vigilance or alertness level of the human. Analyzing this adaptive system with respect to performance and safety margins in a physical robot experiment, we found the proposed method could improve both collaboration efficiency and safety distance.

Low-cost fabrication of flexible tactile sensor arrays.

While for vision and audio the same mass-produced units can be embedded in many different systems from smartphones to robots, tactile sensors have to be built in application-specific shapes and sizes. To use a commercially available tactile sensor, it can be necessary to develop the entire system around an existing sensor model. We present a set of open-source solutions for designing, manufacturing, reading and integrating custom application-specific tactile matrix sensors. Our manufacturing process only requires an off-the-shelf cutting plotter and widely available plastic and metal foils. This allows creating sensors of diverse sizes, shapes, and layouts, which can be adapted to various specific use cases as demonstrated with exemplary robot integrations. For interfacing and readout, we develop an Arduino-like prototype board (Tacduino) with amplifier circuits to ensure good resolution and to suppress crosstalk. As an example, we give step-by-step instructions to build tactile fingertips for the RobotiQ 3-Finger Gripper, and we provide design files for the readout circuit board together with Arduino firmware and driver software. Both, wired and wireless communication between the sensors and a host PC are supported by this system. The hardware was originally presented and investigated in [1].

A Dexterous Hand-Arm Teleoperation System Based on Hand Pose Estimation and Active Vision.

Markerless vision-based teleoperation that leverages innovations in computer vision offers the advantages of allowing natural and noninvasive finger motions for multifingered robot hands. However, current pose estimation methods still face inaccuracy issues due to the self-occlusion of the fingers. Herein, we develop a novel vision-based hand-arm teleoperation system that captures the human hands from the best viewpoint and at a suitable distance. This teleoperation system consists of an end-to-end hand pose regression network and a controlled active vision system. The end-to-end pose regression network (Transteleop), combined with an auxiliary reconstruction loss function, captures the human hand through a low-cost depth camera and predicts joint commands of the robot based on the image-to-image translation method. To obtain the optimal observation of the human hand, an active vision system is implemented by a robot arm at the local site that ensures the high accuracy of the proposed neural network. Human arm motions are simultaneously mapped to the slave robot arm under relative control. Quantitative network evaluation and a variety of complex manipulation tasks, for example, tower building, pouring, and multitable cup stacking, demonstrate the practicality and stability of the proposed teleoperation system.

Reinforcement Learning With Vision-Proprioception Model for Robot Planar Pushing.

We propose a vision-proprioception model for planar object pushing, efficiently integrating all necessary information from the environment. A Variational Autoencoder (VAE) is used to extract compact representations from the task-relevant part of the image. With the real-time robot state obtained easily from the hardware system, we fuse the latent representations from the VAE and the robot end-effector position together as the state of a Markov Decision Process. We use Soft Actor-Critic to train the robot to push different objects from random initial poses to target positions in simulation. Hindsight Experience replay is applied during the training process to improve the sample efficiency. Experiments demonstrate that our algorithm achieves a pushing performance superior to a state-based baseline model that cannot be generalized to a different object and outperforms state-of-the-art policies which operate on raw image observations. At last, we verify that our trained model has a good generalization ability to unseen objects in the real world.

Crossmodal Pattern Discrimination in Humans and Robots: A Visuo-Tactile Case Study.

The quality of crossmodal perception hinges on two factors: The accuracy of the independent unimodal perception and the ability to integrate information from different sensory systems. In humans, the ability for cognitively demanding crossmodal perception diminishes from young to old age. Here, we propose a new approach to research to which degree the different factors contribute to crossmodal processing and the age-related decline by replicating a medical study on visuo-tactile crossmodal pattern discrimination utilizing state-of-the-art tactile sensing technology and artificial neural networks (ANN). We implemented two ANN models to specifically focus on the relevance of early integration of sensory information during the crossmodal processing stream as a mechanism proposed for efficient processing in the human brain. Applying an adaptive staircase procedure, we approached comparable unimodal classification performance for both modalities in the human participants as well as the ANN. This allowed us to compare crossmodal performance between and within the systems, independent of the underlying unimodal processes. Our data show that unimodal classification accuracies of the tactile sensing technology are comparable to humans. For crossmodal discrimination of the ANN the integration of high-level unimodal features on earlier stages of the crossmodal processing stream shows higher accuracies compared to the late integration of independent unimodal classifications. In comparison to humans, the ANN show higher accuracies than older participants in the unimodal as well as the crossmodal condition, but lower accuracies than younger participants in the crossmodal task. Taken together, we can show that state-of-the-art tactile sensing technology is able to perform a complex tactile recognition task at levels comparable to humans. For crossmodal processing, human inspired early sensory integration seems to improve the performance of artificial neural networks. Still, younger participants seem to employ more efficient crossmodal integration mechanisms than modeled in the proposed ANN. Our work demonstrates how collaborative research in neuroscience and embodied artificial neurocognitive models can help to derive models to inform the design of future neurocomputational architectures.