Header logo is

56 results (View BibTeX file of all listed publications)

2017

Thumb xl amd intentiongan
Multi-Modal Imitation Learning from Unstructured Demonstrations using Generative Adversarial Nets

Hausman, K., Chebotar, Y., Schaal, S., Sukhatme, G., Lim, J.

In Proceedings from the conference "Neural Information Processing Systems 2017., (Editors: Guyon I. and Luxburg U.v. and Bengio S. and Wallach H. and Fergus R. and Vishwanathan S. and Garnett R.), Curran Associates, Inc., Advances in Neural Information Processing Systems 30 (NIPS), December 2017 (inproceedings)

am

pdf video [BibTex]

2017

am Hausman, K., Chebotar, Y., Schaal, S., Sukhatme, G., Lim, J. Multi-Modal Imitation Learning from Unstructured Demonstrations using Generative Adversarial Nets In Proceedings from the conference "Neural Information Processing Systems 2017., (Editors: Guyon I. and Luxburg U.v. and Bengio S. and Wallach H. and Fergus R. and Vishwanathan S. and Garnett R.), Curran Associates, Inc., Advances in Neural Information Processing Systems 30 (NIPS), December 2017 (inproceedings)

pdf video [BibTex]

Thumb xl fig1
Active colloidal propulsion over a crystalline surface

Choudhury, U., Straube, A., Fischer, P., Gibbs, J., Höfling, F.

New Journal of Physics, 19, pages: 125010, December 2017 (article)

Abstract
We study both experimentally and theoretically the dynamics of chemically self-propelled Janus colloids moving atop a two-dimensional crystalline surface. The surface is a hexagonally close-packed monolayer of colloidal particles of the same size as the mobile one. The dynamics of the self-propelled colloid reflects the competition between hindered diffusion due to the periodic surface and enhanced diffusion due to active motion. Which contribution dominates depends on the propulsion strength, which can be systematically tuned by changing the concentration of a chemical fuel. The mean-square displacements obtained from the experiment exhibit enhanced diffusion at long lag times. Our experimental data are consistent with a Langevin model for the effectively two-dimensional translational motion of an active Brownian particle in a periodic potential, combining the confining effects of gravity and the crystalline surface with the free rotational diffusion of the colloid. Approximate analytical predictions are made for the mean-square displacement describing the crossover from free Brownian motion at short times to active diffusion at long times. The results are in semi-quantitative agreement with numerical results of a refined Langevin model that treats translational and rotational degrees of freedom on the same footing.

pf

link (url) DOI [BibTex]

pf Choudhury, U., Straube, A., Fischer, P., Gibbs, J., Höfling, F. Active colloidal propulsion over a crystalline surface New Journal of Physics, 19, pages: 125010, December 2017 (article)

link (url) DOI [BibTex]

Thumb xl fig toyex lqr1kernel 1
On the Design of LQR Kernels for Efficient Controller Learning

Marco, A., Hennig, P., Schaal, S., Trimpe, S.

Proceedings of the 56th IEEE Annual Conference on Decision and Control (CDC), pages: 5193-5200, IEEE, IEEE Conference on Decision and Control, December 2017 (conference)

Abstract
Finding optimal feedback controllers for nonlinear dynamic systems from data is hard. Recently, Bayesian optimization (BO) has been proposed as a powerful framework for direct controller tuning from experimental trials. For selecting the next query point and finding the global optimum, BO relies on a probabilistic description of the latent objective function, typically a Gaussian process (GP). As is shown herein, GPs with a common kernel choice can, however, lead to poor learning outcomes on standard quadratic control problems. For a first-order system, we construct two kernels that specifically leverage the structure of the well-known Linear Quadratic Regulator (LQR), yet retain the flexibility of Bayesian nonparametric learning. Simulations of uncertain linear and nonlinear systems demonstrate that the LQR kernels yield superior learning performance.

am ics pn

arXiv PDF On the Design of LQR Kernels for Efficient Controller Learning - CDC presentation DOI Project Page [BibTex]

am ics pn Marco, A., Hennig, P., Schaal, S., Trimpe, S. On the Design of LQR Kernels for Efficient Controller Learning Proceedings of the 56th IEEE Annual Conference on Decision and Control (CDC), pages: 5193-5200, IEEE, IEEE Conference on Decision and Control, December 2017 (conference)

arXiv PDF On the Design of LQR Kernels for Efficient Controller Learning - CDC presentation DOI Project Page [BibTex]

Thumb xl robot legos
Interactive Perception: Leveraging Action in Perception and Perception in Action

Bohg, J., Hausman, K., Sankaran, B., Brock, O., Kragic, D., Schaal, S., Sukhatme, G.

IEEE Transactions on Robotics, 33, pages: 1273-1291, December 2017 (article)

Abstract
Recent approaches in robotics follow the insight that perception is facilitated by interactivity with the environment. These approaches are subsumed under the term of Interactive Perception (IP). We argue that IP provides the following benefits: (i) any type of forceful interaction with the environment creates a new type of informative sensory signal that would otherwise not be present and (ii) any prior knowledge about the nature of the interaction supports the interpretation of the signal. This is facilitated by knowledge of the regularity in the combined space of sensory information and action parameters. The goal of this survey is to postulate this as a principle and collect evidence in support by analyzing and categorizing existing work in this area. We also provide an overview of the most important applications of Interactive Perception. We close this survey by discussing the remaining open questions. Thereby, we hope to define a field and inspire future work.

am

arXiv DOI Project Page [BibTex]

am Bohg, J., Hausman, K., Sankaran, B., Brock, O., Kragic, D., Schaal, S., Sukhatme, G. Interactive Perception: Leveraging Action in Perception and Perception in Action IEEE Transactions on Robotics, 33, pages: 1273-1291, December 2017 (article)

arXiv DOI Project Page [BibTex]

Thumb xl fig1
Wireless Acoustic-Surface Actuators for Miniaturized Endoscopes

Qiu, T., Adams, F., Palagi, S., Melde, K., Mark, A. G., Wetterauer, U., Miernik, A., Fischer, P.

ACS Applied Materials & Interfaces, 9(49):42536 - 42543, November 2017 (article)

Abstract
Endoscopy enables minimally invasive procedures in many medical fields, such as urology. However, current endoscopes are normally cable-driven, which limits their dexterity and makes them hard to miniaturize. Indeed current urological endoscopes have an outer diameter of about 3 mm and still only possess one bending degree of freedom. In this paper, we report a novel wireless actuation mechanism that increases the dexterity and that permits the miniaturization of a urological endoscope. The novel actuator consists of thin active surfaces that can be readily attached to any device and are wirelessly powered by ultrasound. The surfaces consist of two-dimensional arrays of micro-bubbles, which oscillate under ultrasound excitation and thereby generate an acoustic streaming force. Bubbles of different sizes are addressed by their unique resonance frequency, thus multiple degrees of freedom can readily be incorporated. Two active miniaturized devices (with a side length of around 1 mm) are demonstrated: a miniaturized mechanical arm that realizes two degrees of freedom, and a flexible endoscope prototype equipped with a camera at the tip. With the flexible endoscope, an active endoscopic examination is successfully performed in a rabbit bladder. This results show the potential medical applicability of surface actuators wirelessly powered by ultrasound penetrating through biological tissues.

pf

link (url) DOI Project Page [BibTex]

pf Qiu, T., Adams, F., Palagi, S., Melde, K., Mark, A. G., Wetterauer, U., Miernik, A., Fischer, P. Wireless Acoustic-Surface Actuators for Miniaturized Endoscopes ACS Applied Materials & Interfaces, 9(49):42536 - 42543, November 2017 (article)

link (url) DOI Project Page [BibTex]

Thumb xl teaser
Optimizing Long-term Predictions for Model-based Policy Search

Doerr, A., Daniel, C., Nguyen-Tuong, D., Marco, A., Schaal, S., Toussaint, M., Trimpe, S.

Proceedings of 1st Annual Conference on Robot Learning (CoRL), 78, pages: 227-238, (Editors: Sergey Levine and Vincent Vanhoucke and Ken Goldberg), 1st Annual Conference on Robot Learning, November 2017 (conference)

Abstract
We propose a novel long-term optimization criterion to improve the robustness of model-based reinforcement learning in real-world scenarios. Learning a dynamics model to derive a solution promises much greater data-efficiency and reusability compared to model-free alternatives. In practice, however, modelbased RL suffers from various imperfections such as noisy input and output data, delays and unmeasured (latent) states. To achieve higher resilience against such effects, we propose to optimize a generative long-term prediction model directly with respect to the likelihood of observed trajectories as opposed to the common approach of optimizing a dynamics model for one-step-ahead predictions. We evaluate the proposed method on several artificial and real-world benchmark problems and compare it to PILCO, a model-based RL framework, in experiments on a manipulation robot. The results show that the proposed method is competitive compared to state-of-the-art model learning methods. In contrast to these more involved models, our model can directly be employed for policy search and outperforms a baseline method in the robot experiment.

am ics

PDF Project Page [BibTex]

am ics Doerr, A., Daniel, C., Nguyen-Tuong, D., Marco, A., Schaal, S., Toussaint, M., Trimpe, S. Optimizing Long-term Predictions for Model-based Policy Search Proceedings of 1st Annual Conference on Robot Learning (CoRL), 78, pages: 227-238, (Editors: Sergey Levine and Vincent Vanhoucke and Ken Goldberg), 1st Annual Conference on Robot Learning, November 2017 (conference)

PDF Project Page [BibTex]

Thumb xl qg net rev
Acquiring Target Stacking Skills by Goal-Parameterized Deep Reinforcement Learning

Li, W., Bohg, J., Fritz, M.

arXiv, November 2017 (article) Submitted

Abstract
Understanding physical phenomena is a key component of human intelligence and enables physical interaction with previously unseen environments. In this paper, we study how an artificial agent can autonomously acquire this intuition through interaction with the environment. We created a synthetic block stacking environment with physics simulation in which the agent can learn a policy end-to-end through trial and error. Thereby, we bypass to explicitly model physical knowledge within the policy. We are specifically interested in tasks that require the agent to reach a given goal state that may be different for every new trial. To this end, we propose a deep reinforcement learning framework that learns policies which are parametrized by a goal. We validated the model on a toy example navigating in a grid world with different target positions and in a block stacking task with different target structures of the final tower. In contrast to prior work, our policies show better generalization across different goals.

am

arXiv [BibTex]

Thumb xl admi201770108 gra 0001 m
Active Acoustic Surfaces Enable the Propulsion of a Wireless Robot

Qiu, T., Palagi, S., Mark, A. G., Melde, K., Adams, F., Fischer, P.

Advanced Materials Interfaces, 4(21):1700933, September 2017 (article)

Abstract
A major challenge that prevents the miniaturization of mechanically actuated systems is the lack of suitable methods that permit the efficient transfer of power to small scales. Acoustic energy holds great potential, as it is wireless, penetrates deep into biological tissues, and the mechanical vibrations can be directly converted into directional forces. Recently, active acoustic surfaces are developed that consist of 2D arrays of microcavities holding microbubbles that can be excited with an external acoustic field. At resonance, the surfaces give rise to acoustic streaming and thus provide a highly directional propulsive force. Here, this study advances these wireless surface actuators by studying their force output as the size of the bubble-array is increased. In particular, a general method is reported to dramatically improve the propulsive force, demonstrating that the surface actuators are actually able to propel centimeter-scale devices. To prove the flexibility of the functional surfaces as wireless ready-to-attach actuator, a mobile mini-robot capable of propulsion in water along multiple directions is presented. This work paves the way toward effectively exploiting acoustic surfaces as a novel wireless actuation scheme at small scales.

pf

link (url) DOI Project Page [BibTex]

no image
A New Data Source for Inverse Dynamics Learning

Kappler, D., Meier, F., Ratliff, N., Schaal, S.

In Proceedings IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, Piscataway, NJ, USA, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), September 2017 (inproceedings)

am

[BibTex]

am Kappler, D., Meier, F., Ratliff, N., Schaal, S. A New Data Source for Inverse Dynamics Learning In Proceedings IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, Piscataway, NJ, USA, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), September 2017 (inproceedings)

[BibTex]

no image
Bayesian Regression for Artifact Correction in Electroencephalography

Fiebig, K., Jayaram, V., Hesse, T., Blank, A., Peters, J., Grosse-Wentrup, M.

Proceedings of the 7th Graz Brain-Computer Interface Conference 2017 - From Vision to Reality, pages: 131-136, (Editors: Müller-Putz G.R., Steyrl D., Wriessnegger S. C., Scherer R.), Graz University of Technology, Austria, Graz Brain-Computer Interface Conference, September 2017 (conference)

am ei

DOI [BibTex]

am ei Fiebig, K., Jayaram, V., Hesse, T., Blank, A., Peters, J., Grosse-Wentrup, M. Bayesian Regression for Artifact Correction in Electroencephalography Proceedings of the 7th Graz Brain-Computer Interface Conference 2017 - From Vision to Reality, pages: 131-136, (Editors: Müller-Putz G.R., Steyrl D., Wriessnegger S. C., Scherer R.), Graz University of Technology, Austria, Graz Brain-Computer Interface Conference, September 2017 (conference)

DOI [BibTex]

no image
Investigating Music Imagery as a Cognitive Paradigm for Low-Cost Brain-Computer Interfaces

Grossberger, L., Hohmann, M. R., Peters, J., Grosse-Wentrup, M.

Proceedings of the 7th Graz Brain-Computer Interface Conference 2017 - From Vision to Reality, pages: 160-164, (Editors: Müller-Putz G.R., Steyrl D., Wriessnegger S. C., Scherer R.), Graz University of Technology, Austria, Graz Brain-Computer Interface Conference, September 2017 (conference)

am ei

DOI [BibTex]

am ei Grossberger, L., Hohmann, M. R., Peters, J., Grosse-Wentrup, M. Investigating Music Imagery as a Cognitive Paradigm for Low-Cost Brain-Computer Interfaces Proceedings of the 7th Graz Brain-Computer Interface Conference 2017 - From Vision to Reality, pages: 160-164, (Editors: Müller-Putz G.R., Steyrl D., Wriessnegger S. C., Scherer R.), Graz University of Technology, Austria, Graz Brain-Computer Interface Conference, September 2017 (conference)

DOI [BibTex]

Thumb xl jeong et al 2017 advanced science
Corrosion-Protected Hybrid Nanoparticles

Jeong, H. H., Alarcon-Correa, M., Mark, A. G., Son, K., Lee, T., Fischer, P.

Advanced Science, 4(12):1700234, September 2017 (article)

Abstract
Nanoparticles composed of functional materials hold great promise for applications due to their unique electronic, optical, magnetic, and catalytic properties. However, a number of functional materials are not only difficult to fabricate at the nanoscale, but are also chemically unstable in solution. Hence, protecting nanoparticles from corrosion is a major challenge for those applications that require stability in aqueous solutions and biological fluids. Here, this study presents a generic scheme to grow hybrid 3D nanoparticles that are completely encapsulated by a nm thick protective shell. The method consists of vacuum-based growth and protection, and combines oblique physical vapor deposition with atomic layer deposition. It provides wide flexibility in the shape and composition of the nanoparticles, and the environments against which particles are protected. The work demonstrates the approach with multifunctional nanoparticles possessing ferromagnetic, plasmonic, and chiral properties. The present scheme allows nanocolloids, which immediately corrode without protection, to remain functional, at least for a week, in acidic solutions.

pf

link (url) DOI [BibTex]

pf Jeong, H. H., Alarcon-Correa, M., Mark, A. G., Son, K., Lee, T., Fischer, P. Corrosion-Protected Hybrid Nanoparticles Advanced Science, 4(12):1700234, September 2017 (article)

link (url) DOI [BibTex]

Thumb xl screen shot 2017 08 01 at 15.41.10
On the relevance of grasp metrics for predicting grasp success

Rubert, C., Kappler, D., Morales, A., Schaal, S., Bohg, J.

In Proceedings of the IEEE/RSJ International Conference of Intelligent Robots and Systems, September 2017 (inproceedings) Accepted

Abstract
We aim to reliably predict whether a grasp on a known object is successful before it is executed in the real world. There is an entire suite of grasp metrics that has already been developed which rely on precisely known contact points between object and hand. However, it remains unclear whether and how they may be combined into a general purpose grasp stability predictor. In this paper, we analyze these questions by leveraging a large scale database of simulated grasps on a wide variety of objects. For each grasp, we compute the value of seven metrics. Each grasp is annotated by human subjects with ground truth stability labels. Given this data set, we train several classification methods to find out whether there is some underlying, non-trivial structure in the data that is difficult to model manually but can be learned. Quantitative and qualitative results show the complexity of the prediction problem. We found that a good prediction performance critically depends on using a combination of metrics as input features. Furthermore, non-parametric and non-linear classifiers best capture the structure in the data.

am

Project Page [BibTex]

am Rubert, C., Kappler, D., Morales, A., Schaal, S., Bohg, J. On the relevance of grasp metrics for predicting grasp success In Proceedings of the IEEE/RSJ International Conference of Intelligent Robots and Systems, September 2017 (inproceedings) Accepted

Project Page [BibTex]

no image
Local Bayesian Optimization of Motor Skills

Akrour, R., Sorokin, D., Peters, J., Neumann, G.

Proceedings of the 34th International Conference on Machine Learning, 70, pages: 41-50, Proceedings of Machine Learning Research, (Editors: Doina Precup, Yee Whye Teh), PMLR, International Conference on Machine Learning (ICML), August 2017 (conference)

am ei

link (url) Project Page [BibTex]

am ei Akrour, R., Sorokin, D., Peters, J., Neumann, G. Local Bayesian Optimization of Motor Skills Proceedings of the 34th International Conference on Machine Learning, 70, pages: 41-50, Proceedings of Machine Learning Research, (Editors: Doina Precup, Yee Whye Teh), PMLR, International Conference on Machine Learning (ICML), August 2017 (conference)

link (url) Project Page [BibTex]

Thumb xl pilqr cover
Combining Model-Based and Model-Free Updates for Trajectory-Centric Reinforcement Learning

Chebotar, Y., Hausman, K., Zhang, M., Sukhatme, G., Schaal, S., Levine, S.

Proceedings of the 34th International Conference on Machine Learning, 70, Proceedings of Machine Learning Research, (Editors: Doina Precup, Yee Whye Teh), PMLR, International Conference on Machine Learning (ICML), August 2017 (conference)

am

pdf video [BibTex]

am Chebotar, Y., Hausman, K., Zhang, M., Sukhatme, G., Schaal, S., Levine, S. Combining Model-Based and Model-Free Updates for Trajectory-Centric Reinforcement Learning Proceedings of the 34th International Conference on Machine Learning, 70, Proceedings of Machine Learning Research, (Editors: Doina Precup, Yee Whye Teh), PMLR, International Conference on Machine Learning (ICML), August 2017 (conference)

pdf video [BibTex]

Thumb xl fig1
Locomotion of light-driven soft microrobots through a hydrogel via local melting

Palagi, S., Mark, A. G., Melde, K., Qiu, T., Zeng, H., Parmeggiani, C., Martella, D., Wiersma, D. S., Fischer, P.

In 2017 International Conference on Manipulation, Automation and Robotics at Small Scales (MARSS), pages: 1-5, July 2017 (inproceedings)

Abstract
Soft mobile microrobots whose deformation can be directly controlled by an external field can adapt to move in different environments. This is the case for the light-driven microrobots based on liquid-crystal elastomers (LCEs). Here we show that the soft microrobots can move through an agarose hydrogel by means of light-controlled travelling-wave motions. This is achieved by exploiting the inherent rise of the LCE temperature above the melting temperature of the agarose gel, which facilitates penetration of the microrobot through the hydrogel. The locomotion performance is investigated as a function of the travelling-wave parameters, showing that effective propulsion can be obtained by adapting the generated motion to the specific environmental conditions.

pf

DOI [BibTex]

pf Palagi, S., Mark, A. G., Melde, K., Qiu, T., Zeng, H., Parmeggiani, C., Martella, D., Wiersma, D. S., Fischer, P. Locomotion of light-driven soft microrobots through a hydrogel via local melting In 2017 International Conference on Manipulation, Automation and Robotics at Small Scales (MARSS), pages: 1-5, July 2017 (inproceedings)

DOI [BibTex]

no image
Event-based State Estimation: An Emulation-based Approach

Trimpe, S.

IET Control Theory & Applications, 11(11):1684-1693, July 2017 (article)

Abstract
An event-based state estimation approach for reducing communication in a networked control system is proposed. Multiple distributed sensor agents observe a dynamic process and sporadically transmit their measurements to estimator agents over a shared bus network. Local event-triggering protocols ensure that data is transmitted only when necessary to meet a desired estimation accuracy. The event-based design is shown to emulate the performance of a centralised state observer design up to guaranteed bounds, but with reduced communication. The stability results for state estimation are extended to the distributed control system that results when the local estimates are used for feedback control. Results from numerical simulations and hardware experiments illustrate the effectiveness of the proposed approach in reducing network communication.

am ics

arXiv Supplementary material PDF DOI Project Page [BibTex]

am ics Trimpe, S. Event-based State Estimation: An Emulation-based Approach IET Control Theory & Applications, 11(11):1684-1693, July 2017 (article)

arXiv Supplementary material PDF DOI Project Page [BibTex]

Thumb xl comp 3e 00000 copy
Non-Equilibrium Assembly of Light-Activated Colloidal Mixtures

Singh, D. P., Choudhury, U., Fischer, P., Mark, A. G.

Advanced Materials, 29, pages: 1701328, June 2017, 32 (article)

Abstract
The collective phenomena exhibited by artificial active matter systems present novel routes to fabricating out-of-equilibrium microscale assemblies. Here, the crystallization of passive silica colloids into well-controlled 2D assemblies is shown, which is directed by a small number of self-propelled active colloids. The active colloids are titania–silica Janus particles that are propelled when illuminated by UV light. The strength of the attractive interaction and thus the extent of the assembled clusters can be regulated by the light intensity. A remarkably small number of the active colloids is sufficient to induce the assembly of the dynamic crystals. The approach produces rationally designed colloidal clusters and crystals with controllable sizes, shapes, and symmetries. This multicomponent active matter system offers the possibility of obtaining structures and assemblies that cannot be found in equilibrium systems.

pf

link (url) DOI [BibTex]

Thumb xl kim et al 2017 advanced materials
Nanodiamonds That Swim

Kim, J. T., Choudhury, U., Hyeon-Ho, J., Fischer, P.

Advanced Materials, 29(30):1701024, June 2017, Back Cover (article)

Abstract
Nanodiamonds are emerging as nanoscale quantum probes for bio-sensing and imaging. This necessitates the development of new methods to accurately manipulate their position and orientation in aqueous solutions. The realization of an “active” nanodiamond (ND) swimmer in fluids, composed of a ND crystal containing nitrogen vacancy centers and a light-driven self-thermophoretic micromotor, is reported. The swimmer is propelled by a local temperature gradient created by laser illumination on its metal-coated side. Its locomotion—from translational to rotational motion—is successfully controlled by shape-dependent hydrodynamic interactions. The precise engineering of the swimmer's geometry is achieved by self-assembly combined with physical vapor shadow growth. The optical addressability of the suspended ND swimmers is demonstrated by observing the electron spin resonance in the presence of magnetic fields. Active motion at the nanoscale enables new sensing capabilities combined with active transport including, potentially, in living organisms.

pf

link (url) DOI [BibTex]

pf Kim, J. T., Choudhury, U., Hyeon-Ho, J., Fischer, P. Nanodiamonds That Swim Advanced Materials, 29(30):1701024, June 2017, Back Cover (article)

link (url) DOI [BibTex]

Thumb xl apollo system2 croped
Model-Based Policy Search for Automatic Tuning of Multivariate PID Controllers

Doerr, A., Nguyen-Tuong, D., Marco, A., Schaal, S., Trimpe, S.

In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pages: 5295-5301, IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (inproceedings)

am ics

PDF arXiv DOI Project Page [BibTex]

am ics Doerr, A., Nguyen-Tuong, D., Marco, A., Schaal, S., Trimpe, S. Model-Based Policy Search for Automatic Tuning of Multivariate PID Controllers In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pages: 5295-5301, IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (inproceedings)

PDF arXiv DOI Project Page [BibTex]

Thumb xl learning ct block diagram v2
Learning Feedback Terms for Reactive Planning and Control

Rai, A., Sutanto, G., Schaal, S., Meier, F.

Proceedings 2017 IEEE International Conference on Robotics and Automation (ICRA), IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (conference)

am

pdf video [BibTex]

am Rai, A., Sutanto, G., Schaal, S., Meier, F. Learning Feedback Terms for Reactive Planning and Control Proceedings 2017 IEEE International Conference on Robotics and Automation (ICRA), IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (conference)

pdf video [BibTex]

Thumb xl this one
Virtual vs. Real: Trading Off Simulations and Physical Experiments in Reinforcement Learning with Bayesian Optimization

Marco, A., Berkenkamp, F., Hennig, P., Schoellig, A. P., Krause, A., Schaal, S., Trimpe, S.

In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pages: 1557-1563, IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (inproceedings)

am ics pn

PDF arXiv ICRA 2017 Spotlight presentation Virtual vs. Real - Video explanation DOI Project Page [BibTex]

am ics pn Marco, A., Berkenkamp, F., Hennig, P., Schoellig, A. P., Krause, A., Schaal, S., Trimpe, S. Virtual vs. Real: Trading Off Simulations and Physical Experiments in Reinforcement Learning with Bayesian Optimization In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), pages: 1557-1563, IEEE, Piscataway, NJ, USA, IEEE International Conference on Robotics and Automation (ICRA), May 2017 (inproceedings)

PDF arXiv ICRA 2017 Spotlight presentation Virtual vs. Real - Video explanation DOI Project Page [BibTex]

Thumb xl fig quali arm
Probabilistic Articulated Real-Time Tracking for Robot Manipulation

(Best Paper of RA-L 2017, Finalist of Best Robotic Vision Paper Award of ICRA 2017)

Garcia Cifuentes, C., Issac, J., Wüthrich, M., Schaal, S., Bohg, J.

IEEE Robotics and Automation Letters (RA-L), 2(2):577-584, April 2017 (article)

Abstract
We propose a probabilistic filtering method which fuses joint measurements with depth images to yield a precise, real-time estimate of the end-effector pose in the camera frame. This avoids the need for frame transformations when using it in combination with visual object tracking methods. Precision is achieved by modeling and correcting biases in the joint measurements as well as inaccuracies in the robot model, such as poor extrinsic camera calibration. We make our method computationally efficient through a principled combination of Kalman filtering of the joint measurements and asynchronous depth-image updates based on the Coordinate Particle Filter. We quantitatively evaluate our approach on a dataset recorded from a real robotic platform, annotated with ground truth from a motion capture system. We show that our approach is robust and accurate even under challenging conditions such as fast motion, significant and long-term occlusions, and time-varying biases. We release the dataset along with open-source code of our approach to allow for quantitative comparison with alternative approaches.

am

arXiv video code and dataset video PDF DOI Project Page [BibTex]

Thumb xl toc image
Soft 3D-Printed Phantom of the Human Kidney with Collecting System

Adams, F., Qiu, T., Mark, A., Fritz, B., Kramer, L., Schlager, D., Wetterauer, U., Miernik, A., Fischer, P.

Ann. of Biomed. Eng., 45(4):963-972, April 2017 (article)

Abstract
Organ models are used for planning and simulation of operations, developing new surgical instruments, and training purposes. There is a substantial demand for in vitro organ phantoms, especially in urological surgery. Animal models and existing simulator systems poorly mimic the detailed morphology and the physical properties of human organs. In this paper, we report a novel fabrication process to make a human kidney phantom with realistic anatomical structures and physical properties. The detailed anatomical structure was directly acquired from high resolution CT data sets of human cadaveric kidneys. The soft phantoms were constructed using a novel technique that combines 3D wax printing and polymer molding. Anatomical details and material properties of the phantoms were validated in detail by CT scan, ultrasound, and endoscopy. CT reconstruction, ultrasound examination, and endoscopy showed that the designed phantom mimics a real kidney's detailed anatomy and correctly corresponds to the targeted human cadaver's upper urinary tract. Soft materials with a tensile modulus of 0.8-1.5 MPa as well as biocompatible hydrogels were used to mimic human kidney tissues. We developed a method of constructing 3D organ models from medical imaging data using a 3D wax printing and molding process. This method is cost-effective means for obtaining a reproducible and robust model suitable for surgical simulation and training purposes.

pf

DOI Project Page [BibTex]

pf Adams, F., Qiu, T., Mark, A., Fritz, B., Kramer, L., Schlager, D., Wetterauer, U., Miernik, A., Fischer, P. Soft 3D-Printed Phantom of the Human Kidney with Collecting System Ann. of Biomed. Eng., 45(4):963-972, April 2017 (article)

DOI Project Page [BibTex]

Thumb xl fig1
Chapter 8 - Micro- and nanorobots in Newtonian and biological viscoelastic fluids

Palagi, S., (Walker) Schamel, D., Qiu, T., Fischer, P.

In Microbiorobotics, pages: 133 - 162, 8, Micro and Nano Technologies, Second edition, Elsevier, Boston, March 2017 (incollection)

Abstract
Swimming microorganisms are a source of inspiration for small scale robots that are intended to operate in fluidic environments including complex biomedical fluids. Nature has devised swimming strategies that are effective at small scales and at low Reynolds number. These include the rotary corkscrew motion that, for instance, propels a flagellated bacterial cell, as well as the asymmetric beat of appendages that sperm cells or ciliated protozoa use to move through fluids. These mechanisms can overcome the reciprocity that governs the hydrodynamics at small scale. The complex molecular structure of biologically important fluids presents an additional challenge for the effective propulsion of microrobots. In this chapter it is shown how physical and chemical approaches are essential in realizing engineered abiotic micro- and nanorobots that can move in biomedically important environments. Interestingly, we also describe a microswimmer that is effective in biological viscoelastic fluids that does not have a natural analogue.

pf

link (url) DOI [BibTex]

pf Palagi, S., (Walker) Schamel, D., Qiu, T., Fischer, P. Chapter 8 - Micro- and nanorobots in Newtonian and biological viscoelastic fluids In Microbiorobotics, pages: 133 - 162, 8, Micro and Nano Technologies, Second edition, Elsevier, Boston, March 2017 (incollection)

link (url) DOI [BibTex]

Thumb xl eururol2017
Wireless micro-robots for endoscopic applications in urology

Adams, F., Qiu, T., Mark, A. G., Melde, K., Palagi, S., Miernik, A., Fischer, P.

In Eur Urol Suppl, 16(3):e1914, March 2017 (inproceedings)

Abstract
Endoscopy is an essential and common method for both diagnostics and therapy in Urology. Current flexible endoscope is normally cable-driven, thus it is hard to be miniaturized and its reachability is restricted as only one bending section near the tip with one degree of freedom (DoF) is allowed. Recent progresses in micro-robotics offer a unique opportunity for medical inspections in minimally invasive surgery. Micro-robots are active devices that has a feature size smaller than one millimeter and can normally be actuated and controlled wirelessly. Magnetically actuated micro-robots have been demonstrated to propel through biological fluids.Here, we report a novel micro robotic arm, which is actuated wirelessly by ultrasound. It works as a miniaturized endoscope with a side length of ~1 mm, which fits through the 3 Fr. tool channel of a cystoscope, and successfully performs an active cystoscopy in a rabbit bladder.

pf

link (url) DOI [BibTex]

Thumb xl toc image
Pattern formation and collective effects in populations of magnetic microswimmers

Vach, P. J., (Walker) Schamel, D., Fischer, P., Fratzl, P., Faivre, D.

J. of Phys. D: Appl. Phys., 50(11):11LT03, Febuary 2017 (article)

Abstract
Self-propelled particles are one prototype of synthetic active matter used to understand complex biological processes, such as the coordination of movement in bacterial colonies or schools of fishes. Collective patterns such as clusters were observed for such systems, reproducing features of biological organization. However, one limitation of this model is that the synthetic assemblies are made of identical individuals. Here we introduce an active system based on magnetic particles at colloidal scales. We use identical but also randomly-shaped magnetic micropropellers and show that they exhibit dynamic and reversible pattern formation.

pf

DOI [BibTex]

pf Vach, P. J., (Walker) Schamel, D., Fischer, P., Fratzl, P., Faivre, D. Pattern formation and collective effects in populations of magnetic microswimmers J. of Phys. D: Appl. Phys., 50(11):11LT03, Febuary 2017 (article)

DOI [BibTex]

Thumb xl toc image
On-chip enzymatic microbiofuel cell-powered integrated circuits

Mark, A. G., Suraniti, E., Roche, J., Richter, H., Kuhn, A., Mano, N., Fischer, P.

Lab on a Chip, 17(10):1761-1768, Febuary 2017, Recent HOT Article (article)

Abstract
A variety of diagnostic and therapeutic medical technologies rely on long term implantation of an electronic device to monitor or regulate a patient's condition. One proposed approach to powering these devices is to use a biofuel cell to convert the chemical energy from blood nutrients into electrical current to supply the electronics. We present here an enzymatic microbiofuel cell whose electrodes are directly integrated into a digital electronic circuit. Glucose oxidizing and oxygen reducing enzymes are immobilized on microelectrodes of an application specific integrated circuit (ASIC) using redox hydrogels to produce an enzymatic biofuel cell, capable of harvesting electrical power from just a single droplet of 5 mM glucose solution. Optimisation of the fuel cell voltage and power to match the requirements of the electronics allow self-powered operation of the on-board digital circuitry. This study represents a step towards implantable self-powered electronic devices that gather their energy from physiological fluids.

Recent HOT Article.

pf

DOI [BibTex]

pf Mark, A. G., Suraniti, E., Roche, J., Richter, H., Kuhn, A., Mano, N., Fischer, P. On-chip enzymatic microbiofuel cell-powered integrated circuits Lab on a Chip, 17(10):1761-1768, Febuary 2017, Recent HOT Article (article)

DOI [BibTex]

Thumb xl toc image
Strong Rotational Anisotropies Affect Nonlinear Chiral Metamaterials

Hooper, D. C., Mark, A. G., Kuppe, C., Collins, J. T., Fischer, P., Valev, V. K.

Advanced Materials, 29(13):1605110, January 2017 (article)

Abstract
Masked by rotational anisotropies, the nonlinear chiroptical response of a metamaterial is initially completely inaccessible. Upon rotating the sample the chiral information emerges. These results highlight the need for a general method to extract the true chiral contributions to the nonlinear optical signal, which would be hugely valuable in the present context of increasingly complex chiral meta/nanomaterials.

pf

DOI [BibTex]

pf Hooper, D. C., Mark, A. G., Kuppe, C., Collins, J. T., Fischer, P., Valev, V. K. Strong Rotational Anisotropies Affect Nonlinear Chiral Metamaterials Advanced Materials, 29(13):1605110, January 2017 (article)

DOI [BibTex]

no image
Anticipatory Action Selection for Human-Robot Table Tennis

Wang, Z., Boularias, A., Mülling, K., Schölkopf, B., Peters, J.

Artificial Intelligence, 247, pages: 399-414, 2017, Special Issue on AI and Robotics (article)

Abstract
Abstract Anticipation can enhance the capability of a robot in its interaction with humans, where the robot predicts the humans' intention for selecting its own action. We present a novel framework of anticipatory action selection for human-robot interaction, which is capable to handle nonlinear and stochastic human behaviors such as table tennis strokes and allows the robot to choose the optimal action based on prediction of the human partner's intention with uncertainty. The presented framework is generic and can be used in many human-robot interaction scenarios, for example, in navigation and human-robot co-manipulation. In this article, we conduct a case study on human-robot table tennis. Due to the limited amount of time for executing hitting movements, a robot usually needs to initiate its hitting movement before the opponent hits the ball, which requires the robot to be anticipatory based on visual observation of the opponent's movement. Previous work on Intention-Driven Dynamics Models (IDDM) allowed the robot to predict the intended target of the opponent. In this article, we address the problem of action selection and optimal timing for initiating a chosen action by formulating the anticipatory action selection as a Partially Observable Markov Decision Process (POMDP), where the transition and observation are modeled by the \{IDDM\} framework. We present two approaches to anticipatory action selection based on the \{POMDP\} formulation, i.e., a model-free policy learning method based on Least-Squares Policy Iteration (LSPI) that employs the \{IDDM\} for belief updates, and a model-based Monte-Carlo Planning (MCP) method, which benefits from the transition and observation model by the IDDM. Experimental results using real data in a simulated environment show the importance of anticipatory action selection, and that \{POMDPs\} are suitable to formulate the anticipatory action selection problem by taking into account the uncertainties in prediction. We also show that existing algorithms for POMDPs, such as \{LSPI\} and MCP, can be applied to substantially improve the robot's performance in its interaction with humans.

am ei

DOI Project Page [BibTex]

am ei Wang, Z., Boularias, A., Mülling, K., Schölkopf, B., Peters, J. Anticipatory Action Selection for Human-Robot Table Tennis Artificial Intelligence, 247, pages: 399-414, 2017, Special Issue on AI and Robotics (article)

DOI Project Page [BibTex]

no image
Robot Learning

Peters, J., Lee, D., Kober, J., Nguyen-Tuong, D., Bagnell, J., Schaal, S.

In Springer Handbook of Robotics, pages: 357-394, 15, 2nd, (Editors: Siciliano, Bruno and Khatib, Oussama), Springer International Publishing, 2017 (inbook)

am ei

Project Page [BibTex]

am ei Peters, J., Lee, D., Kober, J., Nguyen-Tuong, D., Bagnell, J., Schaal, S. Robot Learning In Springer Handbook of Robotics, pages: 357-394, 15, 2nd, (Editors: Siciliano, Bruno and Khatib, Oussama), Springer International Publishing, 2017 (inbook)

Project Page [BibTex]

2011

no image
Learning, planning, and control for quadruped locomotion over challenging terrain

Kalakrishnan, Mrinal, Buchli, Jonas, Pastor, Peter, Mistry, Michael, Schaal, S.

International Journal of Robotics Research, 30(2):236-258, February 2011 (article)

am

[BibTex]

2011

am Kalakrishnan, Mrinal, Buchli, Jonas, Pastor, Peter, Mistry, Michael, Schaal, S. Learning, planning, and control for quadruped locomotion over challenging terrain International Journal of Robotics Research, 30(2):236-258, February 2011 (article)

[BibTex]

Thumb xl toc image
Quantum-Cascade Laser-Based Vibrational Circular Dichroism

Luedeke, S., Pfeifer, M., Fischer, P.

JOURNAL OF THE AMERICAN CHEMICAL SOCIETY, 133(15):5704-5707, 2011 (article)

Abstract
Vibrational circular dichroism (VCD) spectra were recorded with a tunable external-cavity quantum-cascade laser (QCL). In comparison with standard thermal light sources in the IR, QCLs provide orders of magnitude more power and are therefore promising for VCD studies in strongly absorbing solvents. The brightness of this novel light source is demonstrated with VCD and IR absorption measurements of a number of compounds, including proline in water.

pf

DOI [BibTex]

pf Luedeke, S., Pfeifer, M., Fischer, P. Quantum-Cascade Laser-Based Vibrational Circular Dichroism JOURNAL OF THE AMERICAN CHEMICAL SOCIETY, 133(15):5704-5707, 2011 (article)

DOI [BibTex]

no image
STOMP: Stochastic trajectory optimization for motion planning

Kalakrishnan, M., Chitta, S., Theodorou, E., Pastor, P., Schaal, S.

In IEEE International Conference on Robotics and Automation (ICRA), Shanghai, China, May 9-13, 2011, clmc (inproceedings)

Abstract
We present a new approach to motion planning using a stochastic trajectory optimization framework. The approach relies on generating noisy trajectories to explore the space around an initial (possibly infeasible) trajectory, which are then combined to produced an updated trajectory with lower cost. A cost function based on a combination of obstacle and smoothness cost is optimized in each iteration. No gradient information is required for the particular optimization algorithm that we use and so general costs for which derivatives may not be available (e.g. costs corresponding to constraints and motor torques) can be included in the cost function. We demonstrate the approach both in simulation and on a dual-arm mobile manipulation system for unconstrained and constrained tasks. We experimentally show that the stochastic nature of STOMP allows it to overcome local minima that gradient-based optimizers like CHOMP can get stuck in.

am

link (url) Project Page [BibTex]

am Kalakrishnan, M., Chitta, S., Theodorou, E., Pastor, P., Schaal, S. STOMP: Stochastic trajectory optimization for motion planning In IEEE International Conference on Robotics and Automation (ICRA), Shanghai, China, May 9-13, 2011, clmc (inproceedings)

link (url) Project Page [BibTex]

Thumb xl toc image
Actively coupled cavity ringdown spectroscopy with low-power broadband sources

Petermann, C., Fischer, P.

OPTICS EXPRESS, 19(11):10164-10173, 2011 (article)

Abstract
We demonstrate a coupling scheme for cavity enhanced absorption spectroscopy that makes use of an intracavity acousto-optical modulator to actively switch light into (and out of) a resonator. This allows cavity ringdown spectroscopy (CRDS) to be implemented with broadband nonlaser light sources with spectral power densities of less than 30 mu W/nm. Although the acousto-optical element reduces the ultimate detection limit by introducing additional losses, it permits absorptivities to be measured with a high dynamic range, especially in lossy environments. Absorption measurements for the forbidden transition of gaseous oxygen in air at similar to 760nm are presented using a low-coherence cw-superluminescent diode. The same setup was electronically configured to cover absorption losses from 1.8 x 10(-8)cm(-1) to 7.5\% per roundtrip. This could be of interest in process analytical applications. (C) 2011 Optical Society of America

pf

DOI [BibTex]

pf Petermann, C., Fischer, P. Actively coupled cavity ringdown spectroscopy with low-power broadband sources OPTICS EXPRESS, 19(11):10164-10173, 2011 (article)

DOI [BibTex]

no image
Path Integral Control and Bounded Rationality

Braun, D. A., Ortega, P. A., Theodorou, E., Schaal, S.

In IEEE Symposium on Adaptive Dynamic Programming And Reinforcement Learning (ADPRL), 2011, clmc (inproceedings)

Abstract
Path integral methods [7], [15],[1] have recently been shown to be applicable to a very general class of optimal control problems. Here we examine the path integral formalism from a decision-theoretic point of view, since an optimal controller can always be regarded as an instance of a perfectly rational decision-maker that chooses its actions so as to maximize its expected utility [8]. The problem with perfect rationality is, however, that finding optimal actions is often very difficult due to prohibitive computational resource costs that are not taken into account. In contrast, a bounded rational decision-maker has only limited resources and therefore needs to strike some compromise between the desired utility and the required resource costs [14]. In particular, we suggest an information-theoretic measure of resource costs that can be derived axiomatically [11]. As a consequence we obtain a variational principle for choice probabilities that trades off maximizing a given utility criterion and avoiding resource costs that arise due to deviating from initially given default choice probabilities. The resulting bounded rational policies are in general probabilistic. We show that the solutions found by the path integral formalism are such bounded rational policies. Furthermore, we show that the same formalism generalizes to discrete control problems, leading to linearly solvable bounded rational control policies in the case of Markov systems. Importantly, Bellman?s optimality principle is not presupposed by this variational principle, but it can be derived as a limit case. This suggests that the information- theoretic formalization of bounded rationality might serve as a general principle in control design that unifies a number of recently reported approximate optimal control methods both in the continuous and discrete domain.

am

PDF [BibTex]

am Braun, D. A., Ortega, P. A., Theodorou, E., Schaal, S. Path Integral Control and Bounded Rationality In IEEE Symposium on Adaptive Dynamic Programming And Reinforcement Learning (ADPRL), 2011, clmc (inproceedings)

PDF [BibTex]

no image
Skill learning and task outcome prediction for manipulation

Pastor, P., Kalakrishnan, M., Chitta, S., Theodorou, E., Schaal, S.

In IEEE International Conference on Robotics and Automation (ICRA), Shanghai, China, May 9-13, 2011, clmc (inproceedings)

Abstract
Learning complex motor skills for real world tasks is a hard problem in robotic manipulation that often requires painstaking manual tuning and design by a human expert. In this work, we present a Reinforcement Learning based approach to acquiring new motor skills from demonstration. Our approach allows the robot to learn fine manipulation skills and significantly improve its success rate and skill level starting from a possibly coarse demonstration. Our approach aims to incorporate task domain knowledge, where appropriate, by working in a space consistent with the constraints of a specific task. In addition, we also present an approach to using sensor feedback to learn a predictive model of the task outcome. This allows our system to learn the proprioceptive sensor feedback needed to monitor subsequent executions of the task online and abort execution in the event of predicted failure. We illustrate our approach using two example tasks executed with the PR2 dual-arm robot: a straight and accurate pool stroke and a box flipping task using two chopsticks as tools.

am

link (url) Project Page Project Page [BibTex]

am Pastor, P., Kalakrishnan, M., Chitta, S., Theodorou, E., Schaal, S. Skill learning and task outcome prediction for manipulation In IEEE International Conference on Robotics and Automation (ICRA), Shanghai, China, May 9-13, 2011, clmc (inproceedings)

link (url) Project Page Project Page [BibTex]

no image
An Iterative Path Integral Stochastic Optimal Control Approach for Learning Robotic Tasks

Theodorou, E., Stulp, F., Buchli, J., Schaal, S.

In Proceedings of the 18th World Congress of the International Federation of Automatic Control, 2011, clmc (inproceedings)

Abstract
Recent work on path integral stochastic optimal control theory Theodorou et al. (2010a); Theodorou (2011) has shown promising results in planning and control of nonlinear systems in high dimensional state spaces. The path integral control framework relies on the transformation of the nonlinear Hamilton Jacobi Bellman (HJB) partial differential equation (PDE) into a linear PDE and the approximation of its solution via the use of the Feynman Kac lemma. In this work, we are reviewing the generalized version of path integral stochastic optimal control formalism Theodorou et al. (2010a), used for optimal control and planing of stochastic dynamical systems with state dependent control and diffusion matrices. Moreover we present the iterative path integral control approach, the so called Policy Improvement with Path Integrals or (PI2 ) which is capable of scaling in high dimensional robotic control problems. Furthermore we present a convergence analysis of the proposed algorithm and we apply the proposed framework to a variety of robotic tasks. Finally with the goal to perform locomotion the iterative path integral control is applied for learning nonlinear limit cycle attractors with adjustable land scape.

am

PDF [BibTex]

am Theodorou, E., Stulp, F., Buchli, J., Schaal, S. An Iterative Path Integral Stochastic Optimal Control Approach for Learning Robotic Tasks In Proceedings of the 18th World Congress of the International Federation of Automatic Control, 2011, clmc (inproceedings)

PDF [BibTex]

Thumb xl toc image
Magnetically actuated propulsion at low Reynolds numbers: towards nanoscale control

Fischer, P., Ghosh, A.

NANOSCALE, 3(2):557-563, 2011 (article)

Abstract
Significant progress has been made in the fabrication of micron and sub-micron structures whose motion can be controlled in liquids under ambient conditions. The aim of many of these engineering endeavors is to be able to build and propel an artificial micro-structure that rivals the versatility of biological swimmers of similar size, e. g. motile bacterial cells. Applications for such artificial ``micro-bots'' are envisioned to range from microrheology to targeted drug delivery and microsurgery, and require full motion-control under ambient conditions. In this Mini-Review we discuss the construction, actuation, and operation of several devices that have recently been reported, especially systems that can be controlled by and propelled with homogenous magnetic fields. We describe the fabrication and associated experimental challenges and discuss potential applications.

pf

Video - Nanospropellers DOI [BibTex]

no image
Aerial righting reflexes in flightless animals

Jusufi, A., Zeng, Y., Full, R., Dudley, R.

Integ. Comp. Biol. , 2011 (article)

bio

[BibTex]

bio Jusufi, A., Zeng, Y., Full, R., Dudley, R. Aerial righting reflexes in flightless animals Integ. Comp. Biol. , 2011 (article)

[BibTex]

no image
Bayesian robot system identification with input and output noise

Ting, J., D’Souza, A., Schaal, S.

Neural Networks, 24(1):99-108, 2011, clmc (article)

Abstract
For complex robots such as humanoids, model-based control is highly beneficial for accurate tracking while keeping negative feedback gains low for compliance. However, in such multi degree-of-freedom lightweight systems, conventional identification of rigid body dynamics models using CAD data and actuator models is inaccurate due to unknown nonlinear robot dynamic effects. An alternative method is data-driven parameter estimation, but significant noise in measured and inferred variables affects it adversely. Moreover, standard estimation procedures may give physically inconsistent results due to unmodeled nonlinearities or insufficiently rich data. This paper addresses these problems, proposing a Bayesian system identification technique for linear or piecewise linear systems. Inspired by Factor Analysis regression, we develop a computationally efficient variational Bayesian regression algorithm that is robust to ill-conditioned data, automatically detects relevant features, and identifies input and output noise. We evaluate our approach on rigid body parameter estimation for various robotic systems, achieving an error of up to three times lower than other state-of-the-art machine learning methods

am

link (url) [BibTex]

am Ting, J., D’Souza, A., Schaal, S. Bayesian robot system identification with input and output noise Neural Networks, 24(1):99-108, 2011, clmc (article)

link (url) [BibTex]

no image
Learning variable impedance control

Buchli, J., Stulp, F., Theodorou, E., Schaal, S.

International Journal of Robotics Research, 2011, clmc (article)

Abstract
One of the hallmarks of the performance, versatility, and robustness of biological motor control is the ability to adapt the impedance of the overall biomechanical system to different task requirements and stochastic disturbances. A transfer of this principle to robotics is desirable, for instance to enable robots to work robustly and safely in everyday human environments. It is, however, not trivial to derive variable impedance controllers for practical high degree-of-freedom (DOF) robotic tasks. In this contribution, we accomplish such variable impedance control with the reinforcement learning (RL) algorithm PISq ({f P}olicy {f I}mprovement with {f P}ath {f I}ntegrals). PISq is a model-free, sampling based learning method derived from first principles of stochastic optimal control. The PISq algorithm requires no tuning of algorithmic parameters besides the exploration noise. The designer can thus fully focus on cost function design to specify the task. From the viewpoint of robotics, a particular useful property of PISq is that it can scale to problems of many DOFs, so that reinforcement learning on real robotic systems becomes feasible. We sketch the PISq algorithm and its theoretical properties, and how it is applied to gain scheduling for variable impedance control. We evaluate our approach by presenting results on several simulated and real robots. We consider tasks involving accurate tracking through via-points, and manipulation tasks requiring physical contact with the environment. In these tasks, the optimal strategy requires both tuning of a reference trajectory emph{and} the impedance of the end-effector. The results show that we can use path integral based reinforcement learning not only for planning but also to derive variable gain feedback controllers in realistic scenarios. Thus, the power of variable impedance control is made available to a wide variety of robotic systems and practical applications.

am

link (url) [BibTex]

am Buchli, J., Stulp, F., Theodorou, E., Schaal, S. Learning variable impedance control International Journal of Robotics Research, 2011, clmc (article)

link (url) [BibTex]

no image
Iterative path integral stochastic optimal control: Theory and applications to motor control

Theodorou, E. A.

University of Southern California, University of Southern California, Los Angeles, CA, 2011 (phdthesis)

am

PDF [BibTex]

am Theodorou, E. A. Iterative path integral stochastic optimal control: Theory and applications to motor control University of Southern California, University of Southern California, Los Angeles, CA, 2011 (phdthesis)

PDF [BibTex]

no image
Learning of grasp selection based on shape-templates

Herzog, A.

Karlsruhe Institute of Technology, 2011 (mastersthesis)

am

[BibTex]

am Herzog, A. Learning of grasp selection based on shape-templates Karlsruhe Institute of Technology, 2011 (mastersthesis)

[BibTex]

Thumb xl toc image
Weak value amplified optical activity measurements

Pfeifer, M., Fischer, P.

Opt. Express, 19(17):16508-16517, OSA, 2011 (article)

Abstract
We present a new form of optical activity measurement based on a modified weak value amplification scheme. It has recently been shown experimentally that the left- and right-circular polarization components refract with slightly different angles of refraction at a chiral interface causing a linearly polarized light beam to split into two. By introducing a polarization modulation that does not give rise to a change in the optical rotation it is possible to differentiate between the two circular polarization components even after post-selection with a linear polarizer. We show that such a modified weak value amplification measurement permits the sign of the splitting and thus the handedness of the optically active medium to be determined. Angular beam separations of Δθ ∼ 1 nanoradian, which corresponds to a circular birefringence of Δn ∼ 1 × 10−9, could be measured with a relative error of less than 1%.

pf

link (url) DOI [BibTex]

pf Pfeifer, M., Fischer, P. Weak value amplified optical activity measurements Opt. Express, 19(17):16508-16517, OSA, 2011 (article)

link (url) DOI [BibTex]

no image
Understanding haptics by evolving mechatronic systems

Loeb, G. E., Tsianos, G.A., Fishel, J.A., Wettels, N., Schaal, S.

Progress in Brain Research, 192, pages: 129, 2011 (article)

am

[BibTex]

am Loeb, G. E., Tsianos, G.A., Fishel, J.A., Wettels, N., Schaal, S. Understanding haptics by evolving mechatronic systems Progress in Brain Research, 192, pages: 129, 2011 (article)

[BibTex]

no image
Movement segmentation using a primitive library

Meier, F., Theodorou, E., Stulp, F., Schaal, S.

In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2011), Sept. 25-30, San Francisco, CA, 2011, clmc (inproceedings)

Abstract
Segmenting complex movements into a sequence of primitives remains a difficult problem with many applications in the robotics and vision communities. In this work, we show how the movement segmentation problem can be reduced to a sequential movement recognition problem. To this end, we reformulate the orig-inal Dynamic Movement Primitive (DMP) formulation as a linear dynamical sys-tem with control inputs. Based on this new formulation, we develop an Expecta-tion-Maximization algorithm to estimate the duration and goal position of a par-tially observed trajectory. With the help of this algorithm and the assumption that a library of movement primitives is present, we present a movement seg-mentation framework. We illustrate the usefulness of the new DMP formulation on the two applications of online movement recognition and movement segmen-tation.

am

link (url) [BibTex]

am Meier, F., Theodorou, E., Stulp, F., Schaal, S. Movement segmentation using a primitive library In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2011), Sept. 25-30, San Francisco, CA, 2011, clmc (inproceedings)

link (url) [BibTex]

no image
Learning Force Control Policies for Compliant Manipulation

Kalakrishnan, M., Righetti, L., Pastor, P., Schaal, S.

In 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, pages: 4639-4644, IEEE, San Francisco, USA, sep 2011 (inproceedings)

Abstract
Developing robots capable of fine manipulation skills is of major importance in order to build truly assistive robots. These robots need to be compliant in their actuation and control in order to operate safely in human environments. Manipulation tasks imply complex contact interactions with the external world, and involve reasoning about the forces and torques to be applied. Planning under contact conditions is usually impractical due to computational complexity, and a lack of precise dynamics models of the environment. We present an approach to acquiring manipulation skills on compliant robots through reinforcement learning. The initial position control policy for manipulation is initialized through kinesthetic demonstration. We augment this policy with a force/torque profile to be controlled in combination with the position trajectories. We use the Policy Improvement with Path Integrals (PI2) algorithm to learn these force/torque profiles by optimizing a cost function that measures task success. We demonstrate our approach on the Barrett WAM robot arm equipped with a 6-DOF force/torque sensor on two different manipulation tasks: opening a door with a lever door handle, and picking up a pen off the table. We show that the learnt force control policies allow successful, robust execution of the tasks.

am mg

link (url) DOI [BibTex]

am mg Kalakrishnan, M., Righetti, L., Pastor, P., Schaal, S. Learning Force Control Policies for Compliant Manipulation In 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, pages: 4639-4644, IEEE, San Francisco, USA, sep 2011 (inproceedings)

link (url) DOI [BibTex]