Catching Objects in Flight
Open accessThe DOI matches its Crossref record; 48 of the 52 references it cites were found in a registry or an index; 18 of the 25 sources its claims cite were checked for support.
Abstract
We address the difficult problem of catching in-flight objects with uneven shapes. This requires the solution of three complex problems: accurate prediction of the trajectory of fastmoving objects, predicting the feasible catching configuration, and planning the arm motion, and all within milliseconds. We follow a programming-by-demonstration approach in order to learn, from throwing examples, models of the object dynamics and arm movement. We propose a new methodology to find a feasible catching configuration in a probabilistic manner. We use the dynamical systems approach to encode motion from several demonstrations. This enables a rapid and reactive adaptation of the arm motion in the presence of sensor uncertainty. We validate the approach in simulation with the iCub humanoid robot and in real-world experiments with the KUKA LWR 4+ (7-degree-of-freedom arm robot) to catch a hammer, a tennis racket, an empty bottle, a partially filled bottle, and a cardboard box.
Overview
CoChat Summary
## TL;DR
This paper presents a real-time robot-catching framework that learns object flight dynamics and feasible grasping/reachability from demonstrations, then combines them to predict where and when to catch in-flight objects. It is validated in iCub simulation and on a KUKA LWR 4+ with an Allegro hand, achieving 100% success in simulation and 73.2% in real-world trials.
## Key findings
- The paper reports a 100% success rate for catching two objects in iCub simulation, based on 50 throws; 3 of 50 throws were excluded because they did not enter the robot’s reachable space.
- In real-robot experiments, the system caught 52 out of 71 total trials, yielding a 73.2% success rate; 9 of 80 throws were excluded because they never entered reachable space.
- The system predicts the best catching configuration in about 0.2 ms on a 2.7-GHz quad-core PC, enabling real-time replanning.
- For the real robot, the paper reports mean flying time across all trials of 4.97 ± 0.46 s, and the model predicts trajectories better than a rigid-body model, especially for the partially filled bottle.
## Methods
- The framework has two iterative threads: one continuously predicts object trajectory and updates the best catching configuration and timing; the other adapts arm and finger motion to the updated target.
- Object dynamics are learned from demonstrations using support vector regression with an RBF kernel and coupled with an EKF for noisy sensing; the paper states the model uses about 20 throwing examples per object.
- Graspable space and reachable space are each modeled as Gaussian mixture models, with the optimal catching posture found by maximizing the joint probability of the combined models; arm/hand motion is generated using coupled dynamical systems (CDS) and a timing controller.
- Validation uses iCub simulation with a hammer and tennis racket, and a KUKA LWR 4+ with Allegro hand in real-world experiments on an empty bottle, partially filled bottle, tennis racket, and cardboard box.
## Limitations
- The paper states that the trajectory model is local: if an object is thrown very differently from the demonstrations, prediction can become inaccurate, especially for highly nonlinear motions such as a partially filled bottle.
- The authors say they do not explicitly compute collisions between the object and the hand/arm before catching; they use heuristics instead, and this leads to some failures.
- They state they model only task dynamics, not robot dynamics, so some generated trajectories produce velocities the robot cannot follow.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
14 Valid · 9 Not valid · 2 Not checked · 81 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“Note that this does not require any prior information on the physical properties of the object, such as its mass, position of COM, and moment of inertia.”
What the sources say
- Supportedcites [22] [22] S. Kim and A. Billard, 'Estimating the non-linear dynamics of free-flying objects,' Robot. Auton. Syst. , vol. 60, no. 9, pp. 1108-1122, 2012. - source
Section 2.4 says the DS encoding learns from observed motion without prior information about physical properties such as mass, and that the approach makes no assumption about the center-of-mass position. The passage labeled Table 11 contrasts it with conventional techniques that require properties such as moment of inertia. (LLM, conf=0.94)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“In our previous work [22], we learned from demonstrations the dynamics of more complex objects whose shapes are arbitrary, nonrigid, or their points of interest are not at the COM.”
What the sources say
- Partly supportedcites [22] [22] S. Kim and A. Billard, 'Estimating the non-linear dynamics of free-flying objects,' Robot. Auton. Syst. , vol. 60, no. 9, pp. 1108-1122, 2012. - sourceExperimental — accuracy not yet measured
The Introduction describes learning motion dynamics from human throwing demonstrations, and Section 2.4 says the method handles arbitrarily shaped objects and points away from the center of mass. Section 4.1.1 also studies a half-filled bottle whose moment of inertia changes in flight, but our copy does not establish that the method was applied to nonrigid objects themselves. (LLM, conf=0.84)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“In [38], we show that modeling arm and hand motion through coupled dynamical systems (CDS) is advantageous over the classic approaches reviewed previously.”
What the sources say
- Partly supportedcites [38] [38] A. Shukla and A. Billard, 'Coupled dynamical system based armhand grasping model for learning fast adaptation strategies,' Robot. Auton. Syst. , vol. 60, no. 3, pp. 424-440, 2012. - sourceExperimental — accuracy not yet measured
The Conclusion and discussion says that explicit coupling of hand and finger motion was advantageous over implicit coupling learned with other density-based methods. Section 4.3.1 illustrates that advantage against a single-GMM approach. The paper does not establish an advantage over classic approaches generally, as the citing sentence suggests. (LLM, conf=0.78)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Such a spatial coupling between hand and fingers ensures timely closure of the fingers while following the learned dynamics as closely as possible, even in the presence of arbitrary perturbations.”
What the sources say
- Partly supportedcites [38] [38] A. Shukla and A. Billard, 'Coupled dynamical system based armhand grasping model for learning fast adaptation strategies,' Robot. Auton. Syst. , vol. 60, no. 3, pp. 424-440, 2012. - sourceExperimental — accuracy not yet measured
Section 4.3.1 reports that explicit coupling delays finger closure after a target switch and produces a successful grasp; Section 4.2.2 says finger trajectories resume the unperturbed motion model after perturbation. The claim overextends these results to arbitrary perturbations: Section 5.1 notes that perturbations far from demonstrated motions may yield infeasible robot motions, and the experiments stayed within the center of the robot’s workspace. Page numbers are not available in our copy. (LLM, conf=0.86)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“This offers a robust and fast control of the complete hand-arm system without resorting to extensive parameter tuning.”
What the sources say
- Supportedcites [38] [38] A. Shukla and A. Billard, 'Coupled dynamical system based armhand grasping model for learning fast adaptation strategies,' Robot. Auton. Syst. , vol. 60, no. 3, pp. 424-440, 2012. - source
Section 4.3 reports robust reach-and-grasp control and rapid adaptation to perturbations; Section 4.1 says the implementation controls the arm and finger joints together. Sections 4.2.3 and 5.1 describe just two free parameters that can be inferred from human data and call the parameterization simple and explicit, supporting the claim that extensive tuning is unnecessary—not that no tuning occurs. Page numbers are unavailable in our copy. (LLM, conf=0.86)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
50 more claims · 5 with their cited sources checked
Create a free account to see all 106 claims and every source verdict.
- It may state two separate claims as one. 31
- Probably not a substantial claim. 22
- The quote only partly states this claim. 4
Citations
Cites
Maximum Likelihood from Incomplete Data Via the EM Algorithm
Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1977.tb01600.x · Cited by 50,444
Estimating the Dimension of a Model
Schwarz, G. (1978) · The Annals of Statistics · doi:10.1214/aos/1176344136 · Cited by 47,033
LIBSVM: A library for support vector machines
Chang, C.-C., & Lin, C.-J. (2011) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/1961189.1961199 · Cited by 40,918
On Learning, Representing, and Generalizing a Task in a Humanoid Robot
Calinon, S., Guenter, F., & Billard, A. (2007) · IEEE Transactions on Systems, Man and Cybernetics, Part B (Cybernetics) · doi:10.1109/tsmcb.2006.886952 · Cited by 1,090 · cited 1×
“13 come from the sin- gularity of using Euler angles to represent the orientation (note that the real orientation trajectory of the both models are continuous).”
MethodSelf-citationLearning Stable Nonlinear Dynamical Systems With Gaussian Mixture Models
Khansari-Zadeh, S. M., & Billard, A. (2011) · IEEE Transactions on Robotics · doi:10.1109/tro.2011.2159412 · Cited by 771
Self-citationCapturing robot workspace structure: representing robot capabilities
Zacharias, F., Borst, C., & Hirzinger, G. (2007) · 2007 IEEE/RSJ International Conference on Intelligent Robots and Systems · doi:10.1109/iros.2007.4399105 · Cited by 269
A dynamical system approach to realtime obstacle avoidance
Khansari-Zadeh, S. M., & Billard, A. (2012) · Autonomous Robots · doi:10.1007/s10514-012-9287-y · Cited by 264 · cited 1×
“In Section III, we validate the method in simulation by using the iCub simulator, and in a real robot by using the LWR 4+. Section IV concludes with a summary of remaining issues that will be addressed in future work.”
MethodSelf-citationWorkspaces of planar parallel manipulators
Merlet, J.-P., Gosselin, C. M., & Mouly, N. (1998) · Mechanism and Machine Theory · doi:10.1016/s0094-114x(97)00025-6 · Cited by 240
Planning and Control of Robotic Juggling and Catching Tasks
Buehler, M., Koditschek, D. E., & Kindlmann, P. J. (1994) · The International Journal of Robotics Research · doi:10.1177/027836499401300201 · Cited by 227
Kyriakopoulos, K. J., & Saridis, G. N. (2003) · Proceedings. 1988 IEEE International Conference on Robotics and Automation · doi:10.1109/robot.1988.12075 · Cited by 212
Research works printed but not linked: 1
- [34] S. Schaal and D. Sternad, 'Programmable pattern generators,' in Proc. Int. Conf. Comput. Intell. Neurosci. , 1998, pp. 48-51. -
Cited by
Trends and challenges in robot manipulation
Billard, A., & Kragic, D. (2019) · Science · doi:10.1126/science.aat8414 · Cited by 993
Self-citationAn algorithmic perspective on imitation learning
Osa, T., Pajarinen, J., Neumann, G., Bagnell, J. A., Abbeel, P., & Peters, J. (2018) · Foundations and Trends in Robotics · doi:10.1561/2300000053 · Cited by 405
“The CDS approach has been applied to learn the correlation between the arm and fingers [Shukla and Billard, 2012, Kim et al., 2014], or the eye and arm [Lukic et al., 2014].”
BackgroundMethodInfluentialAn Algorithmic Perspective on Imitation Learning
Osa, T., Pajarinen, J., Neumann, G., Bagnell, J. A., Abbeel, P., & Peters, J. (2018) · now publishers, Inc. eBooks · doi:10.1561/9781680834116 · Cited by 368
Tracking Control of Robotic Manipulators With Uncertain Kinematics and Dynamics
Xiao, B., Yin, S., & Kaynak, O. (2016) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2016.2569068 · Cited by 279
“For instance, object grasping [1], fault detection and diagnosis [2], [3], and robotic manipulation in space for on-orbit servicing [4], have been extensively investigated.”
BackgroundAnyGrasp: Robust and Efficient Grasp Perception in Spatial and Temporal Domains
Fang, H.-S., Wang, C., Fang, H., Gou, M., Liu, J., Yan, H., Liu, W., Xie, Y., & Lu, C. (2023) · IEEE Transactions on Robotics · doi:10.1109/tro.2023.3281153 · Cited by 276
“In order to extend to the dynamic scene, classic methods usually require prior information about the objects [29]–[31], or operate on a fixed set of grasp trajectories [32].”
BackgroundBird-inspired dynamic grasping and perching in arboreal environments
Roderick, W. R. T., Cutkosky, M. R., & Lentink, D. (2021) · Science Robotics · doi:10.1126/scirobotics.abj7562 · Cited by 249
“Other grasping mechanisms can pick up different stationary objects (20–30) or catch objects launched at them (28, 31, 32).”
BackgroundKITECH-Hand: A Highly Dexterous and Modularized Robotic Hand
Lee, D., Park, J., Park, S.-W., Baeg, M.-H., & Bae, J. (2016) · IEEE/ASME Transactions on Mechatronics · doi:10.1109/tmech.2016.2634602 · Cited by 167
“and stable grasping capability have contributed to cutting-edge research on robotic manipulation [24]–[27].”
BackgroundRobust Model Predictive Tracking Control for Robot Manipulators With Disturbances
Dai, L., Yu, Y., Zhai, D., Huang, T., & Xia, Y. (2020) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2020.2984986 · Cited by 141
A unified framework for coordinated multi-arm motion planning
Salehian, S. S. M., Figueroa, N., & Billard, A. (2018) · The International Journal of Robotics Research · doi:10.1177/0278364918765952 · Cited by 148
“If we assume that the object trajectory is predictable, one can simply generalize a single-robot arm feasible posture extraction (Kim et al. 2014) for multiple arms to extract the intercept posture.”
BackgroundSelf-citationDynamic grasp and trajectory planning for moving objects
Marturi, N., Kopicki, M., Rastegarpanah, A., Rajasekaran, V., Adjigble, M., Stolkin, R., Leonardis, A., & Bekiroglu, Y. (2019) · Autonomous Robots · doi:10.1007/s10514-018-9799-1 · Cited by 104
Catch & Carry: reusable neural controllers for vision-guided whole-body tasks
Merel, J., Tunyasuvunakool, S., Ahuja, A., Tassa, Y., Hasenclever, L., Pham, V., Erez, T., Wayne, G., & Heess, N. (2020) · ACM Transactions on Graphics · doi:10.1145/3386569.3392474 · Cited by 168
“…is to design controllers that incorporate domain knowledge for some tasks and to then use learning to refine the behavior around this initial, engineered policy – this scheme has been applied to efforts in robotics for tossing objects [Zeng et al. 2019] and catching objects [Kim et al. 2014].”
BackgroundMethodA Dynamical System Approach for Softly Catching a Flying Object: Theory and Experiment
Salehian, S. S. M., Khoramshahi, M., & Billard, A. (2016) · IEEE Transactions on Robotics · doi:10.1109/tro.2016.2536749 · Cited by 132
“The feasible intercept point (ξ (T ∗)) is estimated by the catching point prediction algorithm proposed in [6].”
BackgroundMethodInfluentialSelf-citationLearning Mobile Manipulation through Deep Reinforcement Learning
Wang, C., Zhang, Q., Tian, Q., Li, S., Wang, X., Lane, D., Petillot, Y., & Wang, S. (2020) · Sensors · doi:10.3390/s20030939 · Cited by 99
“In Reference [26], the authors propose a one-shot visual appearance learning method for a mobile manipulator, which can robustly detect specific objects in a scene following an initial segmentation hint from a human user.”
BackgroundInfluentialPhase estimation for fast action recognition and trajectory generation in human–robot collaboration
Maeda, G. J., Ewerton, M., Neumann, G. M., Lioutikov, R., & Peters, J. (2017) · The International Journal of Robotics Research · doi:10.1177/0278364917693927 · Cited by 87
“The measurement or estimation of velocity, for example, by differentiation of a consistent stream of positions, eliminates the ambiguity illustrated in Figure 2 and allows for the realization of online algorithms that cope with very fast dynamics (Kim et al., 2010, 2014).”
BackgroundRobotic Ball Catching with an Eye-in-Hand Single-Camera System
Cigliano, P., Lippiello, V., Ruggiero, F., & Siciliano, B. (2015) · IEEE Transactions on Control Systems Technology · doi:10.1109/tcst.2014.2380175 · Cited by 79
“Finally, other estimators either make use of proper neural networks [30] or follow a programming-by-demonstration approach to find a feasible catching configuration in a probabilistic manner [31].”
BackgroundTrajectory Tracking of Variable Centroid Objects Based on Fusion of Vision and Force Perception
Gao, H., An, H., Lin, W., Yu, X., & Qiu, J. (2023) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2023.3240502 · Cited by 66
Deep reinforcement learning based moving object grasping
Chen, P., & Wei-qing, L. (2021) · Information Sciences · doi:10.1016/j.ins.2021.01.077 · Cited by 72
A Survey of Wheeled Mobile Manipulation: A Decision-Making Perspective
Thakar, S., Srinivasan, S., Al-Hussaini, S., Bhatt, P. M., Rajendran, P. S., Yoon, Y. J., Dhanaraj, N., Malhan, R. K., Schmid, M., Krovi, V., et al. (2022) · Journal of Mechanisms and Robotics · doi:10.1115/1.4054611 · Cited by 83
Song, H., Kim, Y., Yoon, J., et al. (2018) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594005 · Cited by 57
Dong, H., Asadi, E., Sun, G., Prasad, D. K., & Chen, I. (2018) · IEEE Transactions on Robotics · doi:10.1109/tro.2018.2868804 · Cited by 59
“applications as well, including industrial inspection [22], medical diagnosis [23], [24], recognition of traffic signs [25], security [26], autonomous navigation [15], [27], and tracking targets [28], [29].”
BackgroundHuh, J., & Lee, D. D. (2016) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2016.7487116 · Cited by 67
“We introduce the GMMs for fast collision detection in the high dimensional configuration spaces using the incremental EM clustering, whereas other approaches deal with the GMMs for the imitation of demonstration trajectories [5] or exploit the GMMs to model the distribution of feasible positions and orientations for g…”
MethodAn Intelligent Actuator Fault Reconstruction Scheme for Robotic Manipulators
Xiao, B., & Yin, S. (2017) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2017.2647855 · Cited by 51
“Many tasks can be autonomously and repeatedly carried out using one robot [4], while some tasks would need multiple robots rather than single robot [5].”
BackgroundTran, D. T., Truong, H. V. A., & Ahn, K. K. (2021) · International Journal of Precision Engineering and Manufacturing · doi:10.1007/s12541-020-00427-4 · Cited by 50
Millimeter-Level Pick and Peg-in-Hole Task Achieved by Aerial Manipulator
Wang, M., Chen, Z., Guo, K., Yu, X. Y., Zhang, Y. M., Guo, L., & Wang, W. (2023) · IEEE Transactions on Robotics · doi:10.1109/tro.2023.3338956 · Cited by 74
Inverse Kinematics Based Human Mimicking System using Skeletal Tracking Technology
Alibeigi, M., Rabiee, S., & Ahmadabadi, M. N. (2016) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-016-0384-6 · Cited by 47
“[33] modeled the reachable space for KUKA LWR 4+ and the iCub humanoid robot.”
Survey of robotic arm and parameters
Patidar, V., & Tiwari, R. (2016) · International Conference on Computer Communication and Informatics (ICCCI) · doi:10.1109/iccci.2016.7479938 · Cited by 58
“In catching arm [12], [15], [16], [18], [21], [22], [23] and [24] working envelope cover half of sphere.”
BackgroundInfluentialRen, Y., Zhou, Z., Xu, Z., Yang, Y., Zhai, G., Leibold, M., Ni, F., Zhang, Z., BUSS, M. J., & Zheng, Y. (2024) · IEEE Transactions on Robotics · doi:10.1109/tro.2024.3370048 · Cited by 58
Bimanual compliant tactile exploration for grasping unknown objects
Sommer, N. L., Li, M., & Billard, A. G. (2014) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2014.6907804 · Cited by 44
“This method has been proven to be fast and useful for some real time tasks, such as catching an object in flight [19].”
MethodSelf-citationBenchmark for Human-to-Robot Handovers of Unseen Containers With Unknown Filling
Sánchez-Matilla, R., Chatzilygeroudis, K. I., Modas, A., Duarte, N. F., Xompero, A., Frossard, P., Billard, A., & Cavallaro, A. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.2969200 · Cited by 53
“Marker-based approaches [6], [11] rely on motion capture systems to accurately track in 3D the pose of”
BackgroundSelf-citationLim, V., Huang, H., Chen, L. Y., Wang, J., Ichnowski, J., Seita, D., Laskey, M., & Goldberg, K. Y. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9811651 · Cited by 117
“More recent work has introduced robots that can catch items by predicting item trajectories [21], toss arbitrary objects via self-supervised learning [67], and swing items upwards using tactile feedback [56].”
BackgroundEVOLVER: Online Learning and Prediction of Disturbances for Robot Control
Jia, J., Zhang, W., Guo, K., Wang, J., Yu, X. Y., Shi, Y., & Guo, L. (2023) · IEEE Transactions on Robotics · doi:10.1109/tro.2023.3326318 · Cited by 59
An under actuated robotic arm with adjustable stiffness shape memory polymer joints
Firouzeh, A., Salehian, S. S. M., Billard, A. G., et al. (2015) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2015.7139539 · Cited by 41
“We model the density distribution of the tip positions and the joint configurations using GMM models (Pi) [19].”
MethodSelf-citationGaussian Mixture Model for 3-DoF orientations
Kim, S., Haschke, R., & Ritter, H. J. (2016) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2016.10.002 · Cited by 43
Self-citationCoordinated multi-arm motion planning: Reaching for moving objects in the face of uncertainty
Salehian, S. S. M., Figueroa, N., & Billard, A. (2016) · Robotics: Science and Systems · doi:10.15607/rss.2016.xii.019 · Cited by 39
“If the motion of the object is predictable, one can use a model-based prediction approach [13] and find feasible intercept postures by extending a single-robot arm feasible posture extraction algorithm (such as [12]) to Kr-robot arms.”
MethodSelf-citationBatting an in-flight object to the target
Jia, Y.-B., Gardner, M., & Mu, X. (2019) · The International Journal of Robotics Research · doi:10.1177/0278364918817116 · Cited by 43
“In such a situation, we still use (32) and (33) as approximations to the drag and Magnus forces, by assigning to A the average cross section area during the object’s flight.”
BackgroundMethodInfluentialImpedance Control Design Based on Plastic Deformation for a Robotic Arm
Senoo, T., Koike, M., Murakami, K., & Ishikawa, M. (2016) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2016.2587806 · Cited by 39
“training data set of the object trajectory for accurate estimation [25], [26], minimization of the acceleration of the robot for generating smooth motion [27], and image-based visual servoing with an eye-in-hand configuration to keep the object within the field of view [28].”
MethodImpact-Aware Bimanual Catching of Large-Momentum Objects
Yan, L., Stouraitis, T., Moura, J., Xu, W., Gienger, M., & Vijayakumar, S. (2024) · IEEE Transactions on Robotics · doi:10.1109/tro.2024.3381551 · Cited by 48
“These approaches are particularly useful in cases where the dynamics model of the object are not available, e.g. half-filled water bottle [8].”
BackgroundMethodInfluentialGong, Z., Qiu, C., Tao, B., Bai, H., Yin, Z., & Ding, H. (2020) · Science China Technological Sciences · doi:10.1007/s11431-020-1688-2 · Cited by 33
Catch the Ball: Accurate High-Speed Motions for Mobile Manipulators via Inverse Dynamics Learning
Dong, K., Pereida, K., Shkurti, F., & Schoellig, A. P. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341134 · Cited by 49
“Similar simplifications are also made in the ball catching literature and work well in experiments [1], [2], [9], [22].”
ResultCalinon, S., & Lee, D. (2018) · Springer eBooks · doi:10.1007/978-94-007-6046-2_68 · Cited by 47
Real-time trajectory planning based on joint-decoupled optimization in human-robot interaction
Zhang, S., Zanchettin, A. M., Villa, R., & Dai, S. (2019) · Mechanism and Machine Theory · doi:10.1016/j.mechmachtheory.2019.103664 · Cited by 33
Event-based Agile Object Catching with a Quadrupedal Robot
Forrai, B., Miki, T., Gehrig, D., Hutter, M., & Scaramuzza, D. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10161392 · Cited by 48
“Early work on dynamic object catching focused on the case where the object trajectory is fully observable and known and has shown successful implementation on a variety of robotics platforms ranging from robotic arms [7] to drones [8].”
BackgroundLearning to use chopsticks in diverse gripping styles
Yang, Z., Yin, K., & Liu, L. (2022) · ACM Transactions on Graphics · doi:10.1145/3528223.3530057 · Cited by 42
“predict a feasible catching configuration for object catching tasks [Kim et al. 2014].”
BackgroundMethodBillard, A., Albu‐Schäffer, A., Beetz, M., Burgard, W., Corke, P., Ciocarlie, M., Dahiya, R., Kragić, D., Goldberg, K., Nagai, Y., et al. (2025) · Nature Machine Intelligence · doi:10.1038/s42256-025-01050-6 · Cited by 45
Self-citationHigh-Speed, Small-Deformation Catching of Soft Objects Based on Active Vision and Proximity Sensing
Koyama, K., Murakami, K., Senoo, T., Shimojo, M., & Ishikawa, M. (2019) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2019.2891091 · Cited by 25
“[12] realized catching in-flight objects with uneven shapes using a programming-bydemonstration approach.”
Scaling up liquid state machines to predict over address events from dynamic vision sensors
Kaiser, J., Stal, R., Subramoney, A., Roennau, A., & Dillmann, R. (2017) · Bioinspiration & Biomimetics · doi:10.1088/1748-3190/aa7663 · Cited by 27
“It became a crucial step in many problems such as state estimation [3], visual servoing [4, 5], obstacle avoidance [6, 7] and more recently imitation learning [8, 9].”
BackgroundA fast trajectory planning framework with task-priority for space robot
Luo, J., Yu, M., Wang, M., et al. (2018) · Acta Astronautica · doi:10.1016/j.actaastro.2018.09.023 · Cited by 25
Dynamic Movement Primitives for moving goals with temporal scaling adaptation
Koutras, L., & Doulgeri, Z. (2020) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra40945.2020.9196765 · Cited by 24
“Alternatively one could adapt the motion duration as in [16]; this adaptation is based on the utilization of a model for the goal dynamics which predicts the reaching location.”
MethodGOMP-FIT: Grasp-Optimized Motion Planning for Fast Inertial Transport
Ichnowski, J., Avigal, Y., Liu, Y., & Goldberg, K. Y. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9812387 · Cited by 26
“[14] propose a method to rapidly computes motions to catch an object in flight.”
MethodUnderwater 3D Laser Scanners: The Deformation of the Plane
Palomer, A., Ridao, P., Ribas, D., & Forest, J. (2017) · Lecture notes in control and information sciences · doi:10.1007/978-3-319-55372-6_4 · Cited by 24
Tekülve, J., Fois, A., Sandamirskaya, Y., & Schöner, G. (2019) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2019.00095 · Cited by 30
Towards markerless visual servoing of grasping tasks for humanoid robots
Vicente, P., Jamone, L., & Bernardino, A. J. M. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989441 · Cited by 24
“Real-time reaching/grasping tasks in robots are performed, normally, without any feedback control approach [6], [7], [8], [9], [10], mainly to speed-up the reaching process.”
BackgroundQadeer, N., Shah, J. H., Sharif, M. I., et al. (2022) · Sensors · doi:10.3390/s22062113 · Cited by 22
“There are several challenges for tracking of thrown objects such as continuously changing background, aerodynamic stability and dynamic appearance throughout flight [15].”
BackgroundLearning Reachable Manifold and Inverse Mapping for a Redundant Robot manipulator
Kim, S., & Pérez, J. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9561589 · Cited by 30
“In this context, a continuous reachable manifold learning approach [1], [2] is proposed to validate the feasibility of task space motions for a robot manipulator.”
MethodSelf-citationRobust Motion Prediction of a Free-Tumbling Satellite with On-Ground Experimental Validation
Lampariello, R., Mishra, H., Oumer, N. W., et al. (2021) · Journal of Guidance Control and Dynamics · doi:10.2514/1.g005745 · Cited by 21
Design of Hesitation Gestures for Nonverbal Human-Robot Negotiation of Conflicts
Moon, A., Hashmi, M., der Loos, H. F. M. V., Croft, E. A., & Billard, A. (2021) · ACM Transactions on Human-Robot Interaction · doi:10.1145/3418302 · Cited by 22
“This highly reactive nature of DS-based systems has also been showcased in complex real-time obstacle avoidance tasks, as well as tasks that involve catching flying target objects [21, 25].”
BackgroundSelf-citationTarget-referenced Reactive Grasping for Dynamic Objects
Liu, J., Zhang, R., Fang, H.-S., Gou, M., Fang, H., Wang, C., Xu, S., Yan, H., & Lu, C. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00852 · Cited by 31
“[12,20,28] requires prior knowledge of the objects which is not always available in the real-world.”
BackgroundDynamic Modeling of Hand-Object Interactions via Tactile Sensing
Zhang, Q., Li, Y., Luo, Y., Shou, W., Foshey, M., Yan, J., Tenenbaum, J. B., Matusik, W., & Torralba, A. B. (2021) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros51168.2021.9636361 · Cited by 30
“Given that there is still a huge performance gap between the state-ofthe-art dexterous hand robotic system [19], [20], [21], [22] and human capabilities, we ask a human subject to wear the glove and perform various interaction tasks instead of using a robot system, which allows us to study more dexterous and”
Dual Quaternion-Based Visual Servoing for Grasping Moving Objects
De Farias, C., Adjigble, M., Tamadazte, B., Stolkin, R., & Marturi, N. (2021) · 2021 IEEE 17th International Conference on Automation Science and Engineering (CASE) · doi:10.1109/case49439.2021.9551631 · Cited by 18
“Few works have considered visual feedback to correct hand poses while reaching to grasp [7]–[10].”
BackgroundLearning to Pick at Non-Zero-Velocity From Interactive Demonstrations
Mészáros, A., Franzese, G., & Kober, J. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3165531 · Cited by 18
“Another strategy, which has been employed to absorb impacts particularly in catching tasks [6], involves utilising a follow-through behaviour which continues to track the predicted path of an object even after interception [7].”
BackgroundVisual Reaction: Learning to Play Catch With Your Drone
Zeng, K.-H., Mottaghi, R., Weihs, L., & Farhadi, A. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01159 · Cited by 19
“[20] consider the problem of catching in-flight objects with uneven shapes.”
BackgroundDynamic Capture Using a Traplike Soft Gripper With Stiffness Anisotropy
Yang, S., Zhou, Y., Walker, I. D., Yang, C., Branson, D. T., Song, Z., Dai, J. S., & Kang, R. (2022) · IEEE/ASME Transactions on Mechatronics · doi:10.1109/tmech.2022.3219108 · Cited by 16
“[8] presented a synchronous motion planning method to capture irregular dynamic targets through a machine learning algorithm.”
MethodRobust Ballistic Catching: A Hybrid System Stabilization Problem
Schill, M. M., & BUSS, M. J. (2018) · IEEE Transactions on Robotics · doi:10.1109/tro.2018.2868857 · Cited by 19
“Many previous works consider static catching with a gripper [1]–[4], where the goal is to track and predict the object flight trajectory with high precision as quickly as possible.”
BackgroundCalinon, S., & Lee, D. (2017) · Springer eBooks · doi:10.1007/978-94-007-7194-9_68-1 · Cited by 18
Smart Catcher of weighted Objects
Mothukuri, R., Raju, S. H., Sudarsa, D., et al. (2020) · IOP Conference Series Materials Science and Engineering · doi:10.1088/1757-899x/981/2/022002 · Cited by 16
“In studies particularly done in [10],[11][,[12] and [13] describe the various ways and modes of catching the objects by robot and using its arm.”
BackgroundSocial Moments: A Perspective on Interaction for Social Robotics
Durantin, G., Heath, S., & Wiles, J. (2017) · Frontiers in Robotics and AI · doi:10.3389/frobt.2017.00024 · Cited by 17
“For instance, motion capture cameras can achieve lower latencies and higher framerates and are often used for physical applications where timing is important (e.g., catching objects (Kim et al., 2014)).”
BackgroundMethodBall-Catching System Using Image Processing and an Omni-Directional Wheeled Mobile Robot
Kao, S.-T., & Ho, M. (2021) · Sensors · doi:10.3390/s21093208 · Cited by 18
“In [3], the cameras were located behind the robot for catching in-flight objects with uneven shapes.”
BackgroundHumanoid navigation and heavy load transportation in a cluttered environment
Rioux, A., & Suleiman, W. (2015) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2015.7353669 · Cited by 15
“Our main purpose in this paper is to validate the motion planning approach and whole body control scheme, we therefore, as many related research [16], [17], [18], only to cite few, opted for a complete external sensing and localization.”
MethodReference Spreading: Tracking Performance for Impact Trajectories of a 1DoF Setup
Rijnen, M., Saccon, A., & Nijmeijer, H. (2019) · IEEE Transactions on Control Systems Technology · doi:10.1109/tcst.2019.2898953 · Cited by 16
“Motions with planned collisions naturally occur, for example, in robotic locomotion [1] and dynamic pick-and-place tasks [7], [8].”
Fast Trajectory End-Point Prediction with Event Cameras for Reactive Robot Control
Monforte, M., Gava, L., Iacono, M., Glover, A., & Bartolozzi, C. (2023) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw59228.2023.00422 · Cited by 19
“A drawback of EKF methods, however, is that all the predicted steps from the current instant must be recomputed every time a new measure arrives.”
BackgroundMethodTanaka, K., Nishikawa, S., Niiyama, R., & Kuniyoshi, Y. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3076959 · Cited by 19
“presented a method of catching a flying object using machine-learning techniques [6].”
MethodRen, Z., Hu, B., Wang, Z., et al. (2022) · IEEE Transactions on Robotics · doi:10.1109/tro.2022.3207616 · Cited by 15
“However, in [19], human demonstrations were used to model the graspable space around an object, and probabilistic techniques were employed to model the reachable spaces.”
BackgroundMethodInfluentialProcessing and Forecasting the Trajectory of a Thrown Object Measured by the Stereo Vision System
Mironov, K., Vladimirova, I. A., & Pongratz, M. (2015) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2015.09.155 · Cited by 19
“Later Kim et al. (2014) apply this concept for catching airborne objects with robotic manipulator and achieved 73% success rate.”
MethodMobile Manipulation Integrating Enhanced AMCL High-Precision Location and Dynamic Tracking Grasp
Zhou, H., Chou, W., Tuo, W., Rong, Y., & Xu, S. (2020) · Sensors · doi:10.3390/s20226697 · Cited by 16
“This method captures objects on the conveyor belt at a speed of 0.4 m/s. Similarly, Kim [41] developed an Opti-Track system to track the target ball, load by objects, and quickly predict the trajectory of the target and capture it, where the trajectory of the target was learned and trained by machine-learning algorith…”
MethodInfluentialWang, P., Ma, H., Zhang, Y., et al. (2023) · Sensors · doi:10.3390/s23094412 · Cited by 14
“This method can have high accuracy after tracking the target and is often used for the sorting of lightweight materials in the industrial field with high sorting efficiency [18,19].”
MethodTrajectory Generation With Spatio-Temporal Templates Learned From Demonstrations
Yin, X., & Chen, Q. (2016) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2016.2613843 · Cited by 14
“In [13] and [14], a duration control subsystem was incorporated into the GMR system to address duration-crucial situations, such as catching flying objects, at the expense of maintaining the features learned from demonstrations.”
MethodCastanheira, J. P., Vicente, P., Martínez-Cantín, R., Jamone, L., & Bernardino, A. J. M. (2018) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594009 · Cited by 13
“Indeed, many object grasping controllers rely on a complete open-loop approach to achieve the pre-computed optimal grasp [9], [10], [11], [12], [13], where the robot looks once to the scene, computes the pose of the object, and then drives the arm to the grasping pose without sensory feedback.”
BackgroundSemantic Knowledge Representation for Strategic Interactions in Dynamic Situations
Calvo, C., Villacorta-Atienza, J. A., Diez‐Hermano, S., Khoruzhko, M. A., Lobov, S. A., Potapov, I. A., Sánchez‐Jiménez, A., & Makarov, V. A. (2020) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2020.00004 · Cited by 14
“, from a conveyer belt or even catch fast-moving objects (Kim et al., 2014; Nguyen et al., 2016; Bouyarmane et al., 2018; Mason, 2018).”
BackgroundAnthropomorphic Grasping of Complex-Shaped Objects Using Imitation Learning
Yi, J.-B., Kim, J., Kang, T., Song, D., Park, J., & Yi, S. (2022) · Applied Sciences · doi:10.3390/app122412861 · Cited by 16
A learning from demonstration framework for implementation of a feeding task
Ettehadi, N., & Behal, A. (2018) · Encyclopedia with Semantic Computing and Robotic Intelligence · doi:10.1142/s2529737618500016 · Cited by 15
Planar Robot Casting with Real2Sim2Real Self-Supervised LearningPreprint
Lim, V., Huang, H., Chen, L. Y., Wang, J., Ichnowski, J., Seita, D., Laskey, M., & Goldberg, K. Y. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.04814 · Cited by 13
“More recent work has introduced robots that can catch items by predicting item trajectories [21], toss arbitrary objects via self-supervised learning [67], and swing items upwards using tactile feedback [56].”
BackgroundBimanual dynamic grabbing and tossing of objects onto a moving target
Bombile, M., & Billard, A. (2023) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2023.104481 · Cited by 15
Self-citationGardner, M., & Jia, Y.-B. (2022) · IEEE Transactions on Robotics · doi:10.1109/tro.2022.3165367 · Cited by 13
“Existing methods fail to determine a catching configuration for objects thrown along trajectories that deviate from those used to build machine learned models [8].”
BackgroundMethodFedoseev, A., Serpiva, V., Karmanova, E., Cabrera, M. A., Shirokun, V., Vasilev, I., Savushkin, S., & Tsetserukou, D. (2021) · IEEE International Conference on Soft Robotics (RoboSoft) · doi:10.1109/robosoft51838.2021.9479353 · Cited by 17
Neural Motion Prediction for In-flight Uneven Object Catching
Yu, H., Guo, D., Yin, H., Chen, A., Xu, K., Chen, Z., Wang, M., Tan, Q., Wang, Y., & Xiong, R. (2021) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros51168.2021.9635983 · Cited by 18
“9, the proposed NAE and NAE-DF are superior to Newton’s method and SVR in generalization performance for unseen objects.”
BackgroundMethodInfluentialMaeda, G. J., Neumann, G. M., Ewerton, M., Lioutikov, R., & Peters, J. (2017) · Springer proceedings in advanced robotics · doi:10.1007/978-3-319-60916-4_15 · Cited by 13
“1(b) and allows for the realization of online algorithms that cope with very fast dynamics [10, 11].”
BackgroundImpact Absorbing and Compensation for Heavy Object Catching With an Unmanned Aerial Manipulator
Wang, S., Ma, Z., Quan, F., & Chen, H. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3368300 · Cited by 13
Limb Movement in Dynamic Situations Based on Generalized Cognitive Maps
Villacorta-Atienza, J. A., Calvo, C., Lobov, S. A., & Makarov, V. A. (2017) · Mathematical Modelling of Natural Phenomena · doi:10.1051/mmnp/201712403 · Cited by 10
“It was shown that the robot can make a global prediction and catch an irregular object in flight [14].”
BackgroundMulti-View Shape Estimation of Transparent Containers
Xompero, A., Sánchez-Matilla, R., Modas, A., Frossard, P., & Cavallaro, A. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9054112 · Cited by 14
“Existing methods for localising objects in 3D or estimating their 6 DoF pose rely on databases of 3D object models [5, 6, 7, 8] or need motion capture systems with markers [4, 9, 10].”
MethodImproving reinforcement learning based moving object grasping with trajectory prediction
Xu, B., Hassan, T., & Hussain, I. (2023) · Intelligent Service Robotics · doi:10.1007/s11370-023-00491-5 · Cited by 12
Adaptive Grasping of Moving Objects through Tactile Sensing
Lynch, P., Cullinan, M. F., & McGinn, C. (2021) · Sensors · doi:10.3390/s21248339 · Cited by 15
“the object tracking data from surrounding cameras [2,3,6,25,31].”
BackgroundInfluentialMotion planning for catching a light-weight ball with high-speed visual feedback
Murakami, K., Yamakawa, Y., Senoo, T., & Ishikawa, M. (2015) · IEEE International Conference on Robotics and Biomimetics (ROBIO) · doi:10.1109/robio.2015.7418790 · Cited by 12
“On the other hand, robotic catching is a task with a long history of research, and there have been numerous studies[8], [9], [10].”
BackgroundFast kNN-based prediction for the trajectory of a thrown body
Mironov, K., & Pongratz, M. (2016) · 2006 14th Mediterranean Conference on Control and Automation · doi:10.1109/med.2016.7536007 · Cited by 10
“It was introduced in [1] and then considered in a number of scientific works [2-6].”
BackgroundBall catching with omni-directional wheeled mobile robot and active stereo vision
Kao, S.-T., Wang, Y., & Ho, M. (2017) · IEEE International Symposium on Industrial Electronics (ISIE) · doi:10.1109/isie.2017.8001395 · Cited by 11
“[5-8, 12] used a nonlinear dynamic model with air drag in conjunction with different estimation algorithms to predict the trajectory of a flying ball.”
MethodYu, M., Luo, J., Wang, M., Li, J., Liu, C., & Sun, J. (2022) · Acta Astronautica · doi:10.1016/j.actaastro.2022.11.043 · Cited by 11
Perception for Everyday Human Robot Interaction
Worch, J.-H., Bálint-Benczédi, F., & Beetz, M. (2015) · Künstliche Intell. · doi:10.1007/s13218-015-0400-1 · Cited by 13
“In recent years imitation learning has been the center of attention and remarkable results have been achieved when it comes to reproducing the motions of humans [12, 13].”
BackgroundCalinon, S., & Lee, D. (2018) · Springer eBooks · doi:10.1007/978-94-007-7194-9_68-2 · Cited by 9
Visual Servoing Control Based on EGM Interface of an ABB Robot
Mao, Y., Lu, Q., & Xu, Q. L. (2018) · Chinese Automation Congress · doi:10.1109/cac.2018.8623638 · Cited by 8
“*Qing Xi is an intern of ABB Cooperate Research Center fast moving object [13], [14].”
BackgroundAccuracy of positioning spherical objects with a stereo camera system
Pongratz, M., & Mironov, K. (2015) · IEEE International Conference on Industrial Technology (ICIT) · doi:10.1109/icit.2015.7125326 · Cited by 8
“Robotic catching was introduced more than twenty years ago in [10] and then developed by number of research groups ([11], [12], [13]).”
BackgroundTransport by robotic throwing and catching: Accurate stereo tracking of the spherical object
Mironov, K. (2017) · International Conference on Industrial Engineering, Applications and Manufacturing (ICIEAM) · doi:10.1109/icieam.2017.8076490 · Cited by 8
“[2-7].”
Metadata
- Authors
- Kim, S., Shukla, A., & Billard, A.
- Affiliations
- École Polytechnique Fédérale de Lausanne (CH)
- ORCID
- Billard, Aude
- Venue
- IEEE Transactions on Robotics
- Published
- 2014-10
- Volume
- 30
- Issue
- 5
- Pages
- 1049-1065
- ISSN
- 1552-3098 (ISSN-L)
- 1941-0468
- OpenAlex
- W2018020788
- Cited by
- 220
- References
- 52
- Citations by year
- 2026: 8 · 2025: 14 · 2024: 27 · 2023: 19 · 2022: 20 · 2021: 29 · 2020: 20 · 2019: 12 · 2018: 23 · 2017: 17 · 2016: 18 · 2015: 10 · 2014: 3
- FWCI
- 17.81
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- IEEE Transactions on Robotics (publishedVersion)
- Infoscience (Ecole Polytechnique Fédérale de Lausanne) (submittedVersion)
- Infoscience (Ecole Polytechnique Fédérale de Lausanne) (submittedVersion, cc-by-nc-nd)
- Funders
- EU Projects AMARSI (FP7-ICT-248311)
- FirstMM (FP7-248258)
- Topics
- Physical Sciences › Engineering › Control and Systems Engineering › Robot Manipulation and Learning
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Motor Control and Adaptation
- Keywords
- trajectory prediction, programming by demonstration, dynamical systems, sensor uncertainty, humanoid robots, robot manipulators, iCub
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.