Looking for publications? All our latest publications are listed here, and you can also use our search functions to help find what you are looking for. Power users might also want to consider searching on the EPFL Infoscience site which provides advanced publication search capabilities.

A reconfiguration strategy for modular robots using origami folding

Authors: Yao, Meibao; Belke, Christoph; Cui, Hutao; Paik, Jamie

 

Reconfigurability in versatile systems of modular robots is achieved by changing the morphology of the overall structure as well as by connecting and disconnecting modules. Recurrent connectivity changes can cause misalignment that leads to mechanical failure of the system. This paper presents a new approach to reconfiguration, inspired by the art of origami, that eliminates connectivity changes during transformation. Our method consists of an energy-optimal reconfiguration planner that generates an initial 2D assembly pattern and an actuation sequence of the modular units, both resulting in minimum energy consumption. The algorithmic framework includes two approaches, an automatic modeling algorithm as well as a heuristic algorithm. We further demonstrate the effectiveness of our method by applying the algorithms to Mori, a modular origami robot, in simulation. Our results show that the heuristic algorithm yields reconfiguration schemes with high quality, compared with the automatic modeling algorithm, simultaneously saving a considerable amount of computational time and effort.

Reference

Posted on: June 4, 2019

Unsupervised Moving Object Detection viaContextual Information Separation

Authors: Yang, Yanchao; Loquercio, Antonio; Scaramuzza, Davide; Soatto, Stefano

 

We propose an adversarial contextual model for detecting moving objects in images. A deep neural network istrained to predict the optical flow in a region using information from everywhere else but that region (context), while another network attempts to make such context as uninformative as possible. The result is a model where hypotheses naturally compete with no need for explicit regularization or hyper-parameter tuning. Although our method requires no supervision whatsoever, it outperforms several methods that are pre-trained on large annotated datasets. Our model can be thought of as a generalization of classical variational generative region-based segmentation, but in a way that avoids explicit regularization or solution of partial differential equations at run-time. We publicly release all our code and trained networks

Reference

  • Presented at: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, USA, 2019
  • Read paper
  • Data set
  • Date: 2019
Posted on: May 31, 2019

CED: Color Event Camera Dataset

Authors: Scheerlinck, Cedric; Rebecq, Henri; Stoffregen, Timo; Barnes, Nick; Mahony, Robert; Scaramuzza, Davide

 

Event cameras are novel, bio-inspired visual sensors,whose pixels output asynchronous and independent times-tamped spikes at local intensity changes, called ‘events’. Event cameras offer advantages over conventional frame-based cameras in terms of latency, high dynamic range(HDR) and temporal resolution. Until recently, event cam-eras have been limited to outputting events in the intensity channel, however, recent advances have resulted in the development of color event cameras, such as the Color-DAVIS346. In this work, we present and release the first Color Event Camera Dataset (CED), containing 50 minutes of footage with both color frames and events. CED features a wide variety of indoor and outdoor scenes, which we hope will help drive forward event-based vision research.We also present an extension of the event camera simulator ESIM [1] that enables simulation of color events. Finally,we present an evaluation of three state-of-the-art image re-construction methods that can be used to convert the Color-DAVIS346 into a continuous-time, HDR, color video cam-era to visualise the event stream, and for use in downstream vision applications

Reference

  • Presented at: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, USA
  • Read paper
  • Data set
  • Date: 2019
Posted on: May 31, 2019

Events-to-Video: Bringing Modern Computer Vision to Event Cameras

Authors: Rebecq, Henri; Ranftl, René; Koltun, Vladen; Scaramuzza, Davide

 

Event cameras are novel sensors that report brightnesschanges in the form of asynchronous “events” instead ofintensity frames. They have significant advantages overconventional cameras: high temporal resolution, high dy-namic range, and no motion blur. Since the output of eventcameras is fundamentally different from conventional cam-eras, it is commonly accepted that they require the devel-opment of specialized algorithms to accommodate the par-ticular nature of events. In this work, we take a differ-ent view and propose to apply existing, mature computervision techniques to videos reconstructed from event data.We propose a novel recurrent network to reconstruct videosfrom a stream of events, and train it on a large amountof simulated event data. Our experiments show that ourapproach surpasses state-of-the-art reconstruction meth-ods by a large margin (>20%) in terms of image qual-ity. We further apply off-the-shelf computer vision algo-rithms to videos reconstructed from event data on taskssuch as object classification and visual-inertial odometry,and show that this strategy consistently outperforms algo-rithms that were specifically designed for event data. Webelieve that our approach opens the door to bringing theoutstanding properties of event cameras to an entirely newrange of tasks. A video of the experiments is available at https://www.youtube.com/watch?v=IdYrC4cUO0I

Reference

  • Presented at: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, USA, 2019
  • Read paper
  • Date: 2019
Posted on: May 31, 2019

Harnessing the Rheological Properties of Liquid Metals To Shape Soft Electronic Conductors for Wearable Applications

Authors: Hirsch, Arthur; Dejace, Laurent; Michaud, Hadrien O.; Lacour, Stéphanie P.

 

Emerging applications of the Internet of Things in healthcare, wellness, and gaming require continuous monitoring of the body and its environment, fueling the need for wearable devices able to maintain intimate, reliable, and unobtrusive contact with the human body. This translates in the necessity to develop soft and deformable electronics that match the body’s mechanics and dynamics. In recent years, various strategies have been proposed to form stretchable circuits and more specifically elastic electrical conductors embedded in elastomeric substrate using either geometrical structuring of solid conductors or intrinsically stretchable materials. Gallium (Ga)-based liquid metals (LMs) are an emerging class of materials offering a particularly interesting set of properties for the design of intrinsically deformable conductors. They concomitantly offer the high electrical conductivity of metals with the ability of liquids to flow and reconfigure. The specific chemical and physical properties of Ga-based LMs differ fundamentally from those of solid conductors and need to be considered to successfully process and implement them into stretchable electronic devices. In this Account, we report on how the key physical and chemical properties of Ga-based LMs can be leveraged to enable repeatable manufacturing and precise patterning of stretchable LM conductors. A comprehensive understanding of the interplay between the LM, its receiving substrate chemistry and topography, and the environmental conditions is necessary to meet the reproducibility and reliability standards for large scale deployment in next-generation wearable systems. In oxidative environments, a solid oxide skin forms at the surface of the LM and provides enough stiffness to counterbalance surface tension, and prevent the LM from beading up to a spherical shape. We review techniques that advantageously harness the oxide skin to form metastable structures such as spraying, 3D printing, or channel injection. Next, we explore how controlling the environmental condition prevents the formation or removes the oxide skin, thereby allowing for selective wetting of Ga lyophilic surfaces. Representative examples include selective plating and physical vapor deposition. The wettability of LMs can be further tuned by engineering the surface chemistry and topology of the receiving substrate to form superlyophobic or superlyophilic surfaces. In particular, our group developed Ga-superlyophilic substrates by engineering the surface of silicone rubber with microstructures and a gold coating layer. Thermal evaporation of Ga on such engineered substrates allows for the formation of smooth LM films with micrometric thickness control and design freedom. The versatility of the available deposition techniques facilitates the implementation of LM conductors in a wide variety of wearable devices. We review various epidermal electronic systems using LM conductors as interconnects to carry power and information, transducers and sensors, antennas, and complex hybrid (soft-rigid) electronic circuits. In addition, we highlight the limitations and challenges inherent to the use of Ga LM conductors that include electromigration, corrosion, solidification, and biocompatibility.

Reference

Posted on: May 31, 2019