Capture Point: A Step toward Humanoid Push Recovery

By Jerry E. Pratt, John Carff, Sergey V. Drakunov, Ambarish Goswami. IEEE-RAS International Conference on Humanoid Robots, 2006

1,172 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Civil and Structural Engineering

View Paper: https://www.semanticscholar.org/paper/9d9aabdfd5b47862bdba0b9bb35711065b673418

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Civil and Structural Engineering

Citations in the corpus

Valkyrie: NASA's First Bipedal Humanoid Robot (2015, methodology): “In particular, using WBC to implement a more sophisticated, feedback-driven locomotion strategy, such as capture-point walking (Pratt et al., 2006) will be investigated.”
Prediction and Control with Temporal Segment Models (2017, background): “For example, there are effective and simple manually-designed control laws (Raibert, 1986), (Pratt et al., 2006) that formulate optimal actions as a function of the entire future trajectory rather than a single future state.”
No falls, no resets: Reliable humanoid behavior in the DARPA robotics challenge (2015, methodology): “We compute a “corrected capture point” (CCP) [15], which is an offset to the current capture point.”
Bipedal walking control based on Capture Point dynamics (2011, background): “[8] and Hof [12] independently introduced the Capture Point.”
Dynamic Walking with Compliance on a Cassie Bipedal Robot (2019, methodology): “Results such as [14], [18] designed center of pressure trajectories based on the COM (center of mass) dynamics of the LIP or”
Viability and predictive control for safe locomotion (2008, positive): “For being able to consider at least simple rotational effects such as the bending of the torso which is sometimes helpful for keeping balance, it has been proposed to consider the whole legged robot as a rigid body with constant or varying inertia parameters instead of only a point mass [15], [11].”
Combining model-based policy search with online model learning for control of physical humanoids (2016, background): “To date, a large number of effective approaches to controlling bipedal humanoid robots rely on either execution of pre-created movement trajectories, or manually designed controllers and regulation of control features [2], [5], [11], [20], [21], [24], [25].”
Capturability-based analysis and control of legged locomotion, Part 2: Application to M2V2, a lower-body humanoid (2012, methodology): “Before defining capturability for a legged system we will consider the more general concept of viability of a hybrid dynamic system, as developed in [17] and introduced in the field of legged locomotion in [3], [22].”
The Penn Jerboa: A Platform for Exploring Parallel Composition of Templates (2015, methodology): “The capture point [25] and zero moment point [26] methods use a quasistatic heuristic which is related to these ideas, but are not explicitly designed to servo to desired nonzero speeds.”
A reactive controller framework for quadrupedal locomotion on challenging terrain (2013, methodology): “For example, the 3D Linear Inverted Pendulum [13], the Linear Inverted Pendulum plus Flywheel [14], the Linear Inverted Pendulum with finite-size foot and reactive mass [15].”
MIT Cheetah 3: Design and Control of a Robust, Dynamic Quadruped Robot (2018, methodology): “Each of the footstep locations are calculated from the corresponding hip location using a linear combination of the Raibert heuristic [18] and a velocity-based feedback term from the capture point formulation [19].”
Three-Dimensional Bipedal Walking Control Based on Divergent Component of Motion (2015, background): “Several previous works [13]–[21] propose to split the CoM dynamics into a stable and an unstable part.”
Learning dynamic arm motions for postural recovery (2011, methodology): “[13] introduced the Linear Inverted Pendulum Plus Flywheel model that abstractly models the angular momentum induced by upper body motions as a flywheel about the body center of mass.”
Models, feedback control, and open problems of 3D bipedal robotic walking (2014, background): “[155] considered a ywheel model in order to present the idea of the capture point a point on the ground on which a biped can step and come to a complete (upright) stop without falling over; additional information on capture points can be found in [96, 156].”
Learning agile and dynamic motor skills for legged robots (2019, methodology): “For example, some popular approaches [7–10] use a template-dynamics-based control module that approximates the robot as a point mass with a massless limb to compute the next foothold position.”
Robust physics-based locomotion using low-dimensional planning (2010, methodology): “The Inverted Pendulum Model (IPM), which models center-of-mass (COM) motion during single stance, has been used to control walking [Kajita et al. 2001; Kajita et al. 2003; Pratt et al. 2006; Rebula et al. 2007; Tsai et al. 2010].”
Modular Hopping and Running via Parallel Composition (2017, background): “117)) are exactly the same as of a single “virtual” leg located at the center of mass Raibert (1986), Saranli and Koditschek (2003).”
ANYmal - toward legged robots for harsh environments (2017, methodology): “Thereby, we make use of a simplified inverted pendulum model [32] (similar to the capture point method [33]) in order to balance disturbances from the nominalmotion.”