Efficient and Reactive Planning for High Speed Robot Air Hockey

Liu, Puze; Tateo, Davide; Bou-Ammar, Haitham; Peters, Jan

doi:10.1109/iros51168.2021.9636263

Cited by 9 publications

(21 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…on a real-world Kuka LBR Iiwa 14 robot. This experimental benchmark was previously solved successfully by specialized motion planning algorithms [8], while we show the effectiveness of our general planning method under a strict time budget, and we demonstrate its ability to replan motion on-the-fly. These properties allow for precise dynamic hitting motion, outperforming [8].…”

Section: A Contributionsmentioning

confidence: 78%

“…Planners statistics on the task of rapid movement of a heavy object with orientation constraints and collision avoidance. AQP [8] CBiRRT [11] SST [41] Fig. 5.…”

Section: Mean Vertical Error [Rad] Cnp-b (Ours)mentioning

confidence: 99%

“…using the desired mallet position and goal position, define the desired hitting angle, and add noise to it, iv. compute the desired joint velocity of the robot that maximizes the manipulability along the hitting direction [8] b) Additional loss functions: As we already mentioned in the main text, in this task, to meet the constraints imposed on this task, we introduced an additional constraint manifold loss term, which is responsible to maintain the mallet position on the table surface. We define this loss function as the sum of the losses in x, y, z directions where specific losses are defined by…”

Section: B Experimental Evaluation Detailsmentioning

confidence: 99%

See 2 more Smart Citations

Fast Kinodynamic Planning on the Constraint Manifold with Deep Neural Networks

Kicki¹,

Liu²,

Tateo³

et al. 2023

Preprint

View full text Add to dashboard Cite

Motion planning is a mature area of research in robotics with many well-established methods based on optimization or sampling the state space, suitable for solving kinematic motion planning. However, when dynamic motions under constraints are needed and computation time is limited, fast kinodynamic planning on the constraint manifold is indispensable. In recent years, learning-based solutions have become alternatives to classical approaches, but they still lack comprehensive handling of complex constraints, such as planning on a lowerdimensional manifold of the task space while considering the robot's dynamics. This paper introduces a novel learning-toplan framework that exploits the concept of constraint manifold, including dynamics, and neural planning methods. Our approach generates plans satisfying an arbitrary set of constraints and computes them in a short constant time, namely the inference time of a neural network. This allows the robot to plan and replan reactively, making our approach suitable for dynamic environments. We validate our approach on two simulated tasks and in a demanding real-world scenario, where we use a Kuka LBR Iiwa 14 robotic arm to perform the hitting movement in robotic Air Hockey.

show abstract

Section: A Contributionsmentioning

confidence: 78%

“…Planners statistics on the task of rapid movement of a heavy object with orientation constraints and collision avoidance. AQP [8] CBiRRT [11] SST [41] Fig. 5.…”

Section: Mean Vertical Error [Rad] Cnp-b (Ours)mentioning

confidence: 99%

Section: B Experimental Evaluation Detailsmentioning

confidence: 99%

See 1 more Smart Citation

Fast Kinodynamic Planning on the Constraint Manifold with Deep Neural Networks

Kicki¹,

Liu²,

Tateo³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…reflex [1]) to find the target and react. Solving air hockey with a robot has been achieved using both single [2] and multiple [3], [4] cameras, highlighting that high frame rate sensors (i.e., 120-500 fps) are required to accurately track the puck. The performance benefits from the use of high-speed manipulators, such as KUKA [3] or Franka Emika Panda arm [2] or ad-hoc robots specifically built for the task [4], [5].…”

Section: Introductionmentioning

confidence: 99%

“…Solving air hockey with a robot has been achieved using both single [2] and multiple [3], [4] cameras, highlighting that high frame rate sensors (i.e., 120-500 fps) are required to accurately track the puck. The performance benefits from the use of high-speed manipulators, such as KUKA [3] or Franka Emika Panda arm [2] or ad-hoc robots specifically built for the task [4], [5]. However, it has also been shown that a humanoid robot can play air hockey [5] using its own eyes (cameras) and learning to play from observing a human player and from practice.…”

Section: Introductionmentioning

confidence: 99%

PUCK: Parallel Surface and Convolution-kernel Tracking for Event-Based Cameras

Gava¹,

Monforte²,

Iacono³

et al. 2022

Preprint

View full text Add to dashboard Cite

Low latency and accuracy are fundamental requirements when vision is integrated in robots for high-speed interaction with targets, since they affect system reliability and stability. In such a scenario, the choice of the sensor and algorithms is important for the entire control loop. The technology of event-cameras can guarantee fast visual sensing in dynamic environments, but requires a tracking algorithm that can keep up with the high data rate induced by the robot egomotion while maintaining accuracy and robustness to distractors. In this paper, we introduce a novel tracking method that leverages the Exponential Reduced Ordinal Surface (EROS) data representation to decouple event-by-event processing and tracking computation. The latter is performed using convolution kernels to detect and follow a circular target moving on a plane. To benchmark state-of-the-art event-based tracking, we propose the task of tracking the air hockey puck sliding on a surface, with the future aim of controlling the iCub robot to reach the target precisely and on time. Experimental results demonstrate that our algorithm achieves the best compromise between low latency and tracking accuracy both when the robot is still and when moving.

show abstract