Publications
Below is a list of all relevant publications authored by Robotics Forum members.
Results
- Showing results for:
- Reset all filters
Search results
-
Conference paperAvila Rencoret FB, Mylonas GP, Elson D, 2018,
Robotic Wide-Field Optical Biopsy Imaging For Flexible Endoscopy
, 26th International Congress of the European Association for Endoscopic Surgery (EAES) -
Conference paperTavakoli A, Pardo F, Kormushev P, 2018,
Action branching architectures for deep reinforcement learning
, AAAI 2018, Publisher: AAAIDiscrete-action algorithms have been central to numerousrecent successes of deep reinforcement learning. However,applying these algorithms to high-dimensional action tasksrequires tackling the combinatorial increase of the numberof possible actions with the number of action dimensions.This problem is further exacerbated for continuous-actiontasks that require fine control of actions via discretization.In this paper, we propose a novel neural architecture fea-turing a shared decision module followed by several net-workbranches, one for each action dimension. This approachachieves a linear increase of the number of network outputswith the number of degrees of freedom by allowing a level ofindependence for each individual action dimension. To illus-trate the approach, we present a novel agent, called Branch-ing Dueling Q-Network (BDQ), as a branching variant ofthe Dueling Double Deep Q-Network (Dueling DDQN). Weevaluate the performance of our agent on a set of challeng-ing continuous control tasks. The empirical results show thatthe proposed agent scales gracefully to environments with in-creasing action dimensionality and indicate the significanceof the shared decision module in coordination of the dis-tributed action branches. Furthermore, we show that the pro-posed agent performs competitively against a state-of-the-art continuous control algorithm, Deep Deterministic PolicyGradient (DDPG).
-
Conference paperElson D, Avila Rencoret F, Mylonas G, 2018,
Robotic Wide-Field Optical Biopsy Imaging for Flexible Endoscopy (Gerhard Buess Technology Award)
, 26th Annual International EAES Congress -
Conference paperZhao M, Oude Vrielink T, Elson D, et al., 2018,
Endoscopic TORS-CYCLOPS: A Novel Cable-driven Parallel Robot for Transoral Laser Surgery
, 26th Annual International EAES Congress -
Book chapterPorta JM, Rojas N, Thomas F, 2018,
Distance geometry in active structures
, Mechatronics for Cultural Heritage and Civil Engineering, Editors: Ottaviano, Pelliccio, Gattulli, Publisher: Springer, Pages: 115-136Distance constraints are an emerging formulation that offers intuitive geometrical interpretation of otherwise complex problems. The formulation can be applied in problems such as position and singularity analysis and path planning of mechanisms and structures. This paper reviews the recent advances in distance geometry, providing a unified view of these apparently disparate problems. This survey reviews algebraic and numerical techniques, and is, to the best of our knowledge, the first attempt to summarize the different approaches relating to distance-based formulations.
-
Conference paperKanajar P, Caldwell DG, Kormushev P, 2017,
Climbing over large obstacles with a humanoid robot via multi-contact motion planning
, IEEE RO-MAN 2017: 26th IEEE International Symposium on Robot and Human Interactive Communication, Publisher: IEEE, Pages: 1202-1209Incremental progress in humanoid robot locomotion over the years has achieved important capabilities such as navigation over flat or uneven terrain, stepping over small obstacles and climbing stairs. However, the locomotion research has mostly been limited to using only bipedal gait and only foot contacts with the environment, using the upper body for balancing without considering additional external contacts. As a result, challenging locomotion tasks like climbing over large obstacles relative to the size of the robot have remained unsolved. In this paper, we address this class of open problems with an approach based on multi-body contact motion planning guided through physical human demonstrations. Our goal is to make the humanoid locomotion problem more tractable by taking advantage of objects in the surrounding environment instead of avoiding them. We propose a multi-contact motion planning algorithm for humanoid robot locomotion which exploits the whole-body motion and multi-body contacts including both the upper and lower body limbs. The proposed motion planning algorithm is applied to a challenging task of climbing over a large obstacle. We demonstrate successful execution of the climbing task in simulation using our multi-contact motion planning algorithm initialized via a transfer from real-world human demonstrations of the task and further optimized.
-
Conference paperZhang F, Cully A, Demiris YIANNIS, 2017,
Personalized Robot-assisted Dressing using User Modeling in Latent Spaces
, 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Publisher: IEEE, ISSN: 2153-0866Robots have the potential to provide tremendous support to disabled and elderly people in their everyday tasks, such as dressing. Many recent studies on robotic dressing assistance usually view dressing as a trajectory planning problem. However, the user movements during the dressing process are rarely taken into account, which often leads to the failures of the planned trajectory and may put the user at risk. The main difficulty of taking user movements into account is caused by severe occlusions created by the robot, the user, and the clothes during the dressing process, which prevent vision sensors from accurately detecting the postures of the user in real time. In this paper, we address this problem by introducing an approach that allows the robot to automatically adapt its motion according to the force applied on the robot's gripper caused by user movements. There are two main contributions introduced in this paper: 1) the use of a hierarchical multi-task control strategy to automatically adapt the robot motion and minimize the force applied between the user and the robot caused by user movements; 2) the online update of the dressing trajectory based on the user movement limitations modeled with the Gaussian Process Latent Variable Model in a latent space, and the density information extracted from such latent space. The combination of these two contributions leads to a personalized dressing assistance that can cope with unpredicted user movements during the dressing while constantly minimizing the force that the robot may apply on the user. The experimental results demonstrate that the proposed method allows the Baxter humanoid robot to provide personalized dressing assistance for human users with simulated upper-body impairments.
-
Conference paperRakicevic N, Kormushev P, 2017,
Efficient Robot Task Learning and Transfer via Informed Search in Movement Parameter Space
, Workshop on Acting and Interacting in the Real World: Challenges in Robot Learning, 31st Conference on Neural Information Processing Systems (NIPS 2017) -
Conference paperTavakoli A, Pardo F, Kormushev P, 2017,
Action Branching Architectures for Deep Reinforcement Learning
, Deep Reinforcement Learning Symposium, 31st Conference on Neural Information Processing Systems (NIPS 2017) -
Conference paperChoi J, Chang HJ, Yun S, et al., 2017,
Attentional correlation filter network for adaptive visual tracking
, IEEE Conference on Computer Vision and Pattern Recognition, Publisher: IEEE, ISSN: 1063-6919We propose a new tracking framework with an attentional mechanism that chooses a subset of the associated correlation filters for increased robustness and computational efficiency. The subset of filters is adaptively selected by a deep attentional network according to the dynamic properties of the tracking target. Our contributions are manifold, and are summarised as follows: (i) Introducing the Attentional Correlation Filter Network which allows adaptive tracking of dynamic targets. (ii) Utilising an attentional network which shifts the attention to the best candidate modules, as well as predicting the estimated accuracy of currently inactive modules. (iii) Enlarging the variety of correlation filters which cover target drift, blurriness, occlusion, scale changes, and flexible aspect ratio. (iv) Validating the robustness and efficiency of the attentional mechanism for visual tracking through a number of experiments. Our method achieves similar performance to non real-time trackers, and state-of-the-art performance amongst real-time trackers.
-
Conference paperYoo YJ, Chang H, Yun S, et al., 2017,
Variational autoencoded regression: high dimensional regression of visual data on complex manifold
, IEEE Conference on Computer Vision and Pattern Recognition, Publisher: IEEE, Pages: 2943-2952This paper proposes a new high dimensional regression method by merging Gaussian process regression into a variational autoencoder framework. In contrast to other regression methods, the proposed method focuses on the case where output responses are on a complex high dimensional manifold, such as images. Our contributions are summarized as follows: (i) A new regression method estimating high dimensional image responses, which is not handled by existing regression algorithms, is proposed. (ii) The proposed regression method introduces a strategy to learn the latent space as well as the encoder and decoder so that the result of the regressed response in the latent space coincide with the corresponding response in the data space. (iii) The proposed regression is embedded into a generative model, and the whole procedure is developed by the variational autoencoder framework. We demonstrate the robustness and effectiveness of our method through a number of experiments on various visual data regression problems.
-
Conference paperRojas N, Dollar AM, 2017,
Distance-based kinematics of the five-oblique-axis thumb model with intersecting axes at the metacarpophalangeal joint
, 2017 IEEE RAS/EMBS International Conference on Rehabilitation Robotics, Publisher: IEEE, ISSN: 1945-7901This paper proposes a novel and simple methodto compute all possible solutions of the inverse kinematicsproblem of the five-oblique-axis thumb model with intersectingaxes at the metacarpophalangeal joint. This thumb model isone of the suggested results by a magnetic-resonance-imaging-based study that, in contrast to those based on cadaver fingersor on the tracking of the surface of the fingers, takes intoaccount muscle and ligament behaviors and avoids inaccuraciesresulting from the movement of the skin with respect to thebones. The proposed distance-based inverse kinematics methodeliminates the use of arbitrary reference frames as is usuallyrequired by standard approaches; this is relevant because thenumerical conditioning of the resulting system of equationswith such traditional approaches depends on the selectedreference frames. Moreover, contrary to other parametrizations(e.g., Denavit-Hartenberg parameters), the suggested distance-based parameters for the thumb have a natural, human-understandable geometric meaning that makes them easier tobe determined from any posture. These characteristics makethe proposed approach of interest for those working in, forinstance, measuring and modeling the movement of the humanhand, developing rehabilitation devices such as orthoses andprostheses, or designing anthropomorphic robotic hands.
-
Conference paperBircher WG, Dollar AM, Rojas N, 2017,
A two-fingered robot gripper with large object reorientation range
, 2017 IEEE International Conference on Robotics and Automation (ICRA), Publisher: IEEEIt is very challenging for a robotic gripper to achieve large reorientations with grasped objects without accidental object ejection. This paper presents a simple gripper that can repeatedly achieve large reorientations over π/2 rad through the kinematics of the hand-object system alone, without the use of high fidelity contact sensors, complex control of active finger surfaces, or highly actuated fingers. This gripper is the result of two kinematic parameter search optimizations connected in cascade. Besides the large range of reorientation attained, the obtained gripper also corresponds to a novel topology since ternary joints in the palm are presented. The in-hand planar reorientation capabilities of the proposed gripper are experimentally tested with success.
-
Conference paperKanner O, Rojas N, Dollar AM, 2017,
Between-leg coupling schemes for passively-adaptive non-redundant legged robots
, 2017 IEEE International Conference on Robotics and Automation (ICRA), Publisher: IEEEThis paper studies the synthesis of between-leg coupling schemes for passively-adaptive non-redundant legged robots. Highly actuated legged robots can arbitrarily locate their feet relative to their bodies through active control, but often wind up kinematically over-constrained following ground contact, requiring complex redundant control for stable locomotion. The use of passive sprung joints can provide some minimal passive adaptability to terrain, but it is limited to relatively low terrain variability due to practical travel limits. In this paper, using a 4-RR platform as case study, we show that implementing parallel adaptive couplings between legs of a stance platform can yield substantial passive adaptability to rough terrain while still ensuring that the body is fully constrained in stance. This study uses screw theory-based mobility analysis methods to determine the number of constraints required to control the stance platform. Several coupling schemes are then considered and evaluated through a simulation of their stance capabilities over arbitrary terrain. An experimental validation of these simulation results is presented; it demonstrates the viability of the proposed scheme for passive adaptability.
-
Conference paperRojas N, Thomas F, 2017,
Forward kinematics of the general triple-arm robot using a distance-based formulation
, 7th IFToMM International Workshop on Computational Kinematics, Publisher: SpringerDistance-based formulations have successfully been used to obtain closure polynomialsfor planar mechanisms without relying, in most cases, on variable eliminations. The methods re-sulting from previous attempts to generalize these techniques to spatial mechanisms exhibit somelimitations such as the impossibility of incorporating orientation constraints. For the first time, thispaper presents a complete satisfactory generalization. As an example, it is applied to obtain a clo-sure polynomial for the the general triple-arm parallel robot (that is, the 3-RPS 3-DOF robot). Thispolynomial, not linked to any particular reference frame, is obtained without variable eliminationsor tangent-half-angle substitutions.
-
Journal articleWard-Cherrier B, Rojas N, Lepora NF, 2017,
Model-free precise in-hand manipulation with a 3D-printed tactile gripper
, IEEE Robotics and Automation Letters, Vol: 2, Pages: 2056-2063, ISSN: 2377-3766The use of tactile feedback for precision manipulation in robotics still lags far behind human capabilities. This study has two principal aims: 1) to demonstrate in-hand reorientation of grasped objects through active tactile manipulation; and 2) to present the development of a novel TacTip sensor and a GR2 gripper platform for tactile manipulation. Through the use of Bayesian active perception algorithms, the system successfully achieved inhand reorientation of cylinders of different diameters (20, 25, 30, and 35 mm) using tactile feedback. Average orientation errors along manipulation trajectories were below 5° for all cylinders with reorientation ranges varying from 42° to 67°. We also demonstrated an improvement in active tactile manipulation accuracy when using additional training data. Our methods for active tactile manipulation with the GR2 TacTip gripper are model free, can be used to investigate principles of dexterous manipulation, and could lead to essential advances in the areas of robotic tactile manipulation and teleoperated robots.
-
Journal articleKanner OY, Rojas N, Odhner LU, et al., 2017,
Adaptive legged robots through exactly-constrained and non-redundant design
, IEEE Access, Vol: 5, Pages: 11131-11141, ISSN: 2169-3536This paper presents a novel strategy for designing passively adaptive, statically stable walking robots with full body mobility that are exactly constrained and non-redundantly actuated during stance. In general, fully mobile legged robots include a large number of actuated joints, giving them a wide range of controllable foot placements but resulting in overconstraint during stance, requiring kinematic redundancy and redundant control for effective locomotion. The proposed design strategy allows for the elimination of actuation redundancy, thus greatly reducing the weight and complexity of the legged robots obtained and allowing for simpler control schemes. Moreover, the underconstrained nature of the resulting robots during swing allows for passive adaptability to rough terrain without large contact forces. The strategy uses kinematic mobility analysis tools to synthesize leg topologies, underactuated robotics design approaches to effectively distribute actuation constraints, and elastic elements to influence nominal leg behavior. Several examples of legged robot designs using the suggested approach are thoroughly discussed and a proof-of-concept of a non-redundant walking robot is presented.
-
Journal articleMa RR, Rojas N, Dollar AM, 2016,
Spherical hands: toward underactuated, in-hand manipulation invariant to object size and grasp location
, Journal of Mechanisms and Robotics, Vol: 8, Pages: 061021-061021-12, ISSN: 1942-4302Minimalist, underactuated hand designs can be modified to produce useful, dexterous, in-hand capabilities without sacrificing their passive adaptability in power grasping. Incorporating insight from studies in parallel mechanisms, we implement and investigate the “spherical hand” morphologies: novel, hand topologies with two fingers configured such that the instantaneous screw axes, describing the displacement of the grasped object, always intersect at the same point relative to the palm. This produces the same instantaneous motion about a common point for any object geometry in a stable grasp. Various rotary fingertip designs are also implemented to help maintain stable contact conditions and minimize slip, in order to prove the feasibility of this design in physical hand implementations. The achievable precision manipulation workspaces of the proposed morphologies are evaluated and compared to prior human manipulation data as well as manipulation results with traditional three-finger hand topologies. Experiments suggest that the spherical hands' design modifications can make the system's passive reconfiguration more easily predictable, providing insight into the expected object workspace while minimizing the dependence on accurate object and contact modeling. We believe that this design can significantly reduce the complexity of planning and executing dexterous manipulation movements in unstructured environments with underactuated hands.
-
Journal articlePalomeras N, Carrera A, Hurtós N, et al., 2016,
Toward persistent autonomous intervention in a subsea panel
, Autonomous Robots, Vol: 40, Pages: 1279-1306 -
Journal articleRojas N, Dollar AM, 2016,
Classification and kinematic equivalents of contact types for fingertip-based robot hand manipulation
, Journal of Mechanisms and Robotics, Vol: 8, Pages: 041014-1-041014-9, ISSN: 1942-4302In the context of robot manipulation, Salisbury's taxonomy is the common standard used to define the types of contact interactions that can occur between the robot and a contacted object; the basic concept behind such classification is the modeling of contacts as kinematic pairs. In this paper, we extend this notion by modeling the effects of a robot contacting a body as kinematic chains. The introduced kinematic-chain-based contact model is based on an extension of the Bruyninckx–Hunt approach of surface–surface contact. A general classification of nonfrictional and frictional contact types suitable for both manipulation analyses and robot hand design is then proposed, showing that all standard contact categories used in robotic manipulation are special cases of the suggested generalization. New contact models, such as ball, tubular, planar translation, and frictional adaptive finger contacts, are defined and characterized. An example of manipulation analysis that lays out the relevance and practicality of the proposed classification is detailed.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.