Optimization-based locomotion planning, estimation, and control design for the atlas humanoid robot

By Scott Kuindersma, Robin Deits, Maurice F. Fallon, Andres K. Valenzuela, Hongkai Dai, Frank Permenter, Twan Koolen, Pat Marion, Russ Tedrake. 2015

920 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Biomedical Engineering

How researchers see this paper

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

In Aerospace Engineering

In Biomedical Engineering

Citations in the corpus

Verifiable Reinforcement Learning via Policy Extraction (2018, background): “For example, self-driving cars must robustly handle a variety of human behaviors [26], controllers for robotics typically need stability guarantees [2, 20, 8], and air trafﬁc control should provably satisfy safety properties including robustness [19].”
The limits and potentials of deep learning for robotics (2018, background): “These model-based techniques often capture properties such as the mass, momentum, shape, and surface friction of objects, and use these to generate controls that change the environment in a desirable way [57], [59], [105], [119].”
Reaching the limit in autonomous racing: Optimal control versus reinforcement learning (2023, background): “OC, on the other hand, has been successful primarily in tackling control problems in fairly well-understood dynamical systems [18–22].”
Contact-aided invariant extended Kalman filtering for robot state estimation (2019, methodology): “Consequently, nonlinear observers are typically used to fuse leg odometry and inertial measurements to infer trajectory, velocity, and calibration parameters (Benallegue and Lamiraux, 2015; Eljaik et al., 2015; Kuindersma et al., 2016; Rotella et al., 2014).”
Deep Kernels for Optimizing Locomotion Controllers (2017, background): “However, their controller parametrization is very different, and not widely used, unlike our inverse dynamics and force-based controller which is more modern and state-of-the-art [7], [21], [6].”
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”
Probabilistic Terrain Mapping for Mobile Robots With Uncertain Localization (2018, neutral): “In the context of rough terrain locomotion, many approaches (e.g., [6, 7, 8, 9]) rely on a grid-based, 2.5D height map representation of the terrain, where each cell in the map represents the terrain height at that position.”
HumanPlus: Humanoid Shadowing and Imitation from Humans (2024, background): “While model-based control [16, 36, 40, 55, 69, 92, 98] has made tremendous progress on a wide variety of humanoid robots [13, 32, 37, 54, 65, 89], learning-based meth-ods can achieve robust locomotion performance for humanoids [39, 67, 68, 85, 91, 94, 103] and biped robots [4, 42, 44, 45, 83, 84,…”
Data Efficient Reinforcement Learning for Legged Robots (2019, background): “Robust and agile locomotion of legged robots based on classical control stacks typically requires accurate dynamics models, human expertise, and tedious manual tuning [1, 2, 3, 4].”
Modular Hopping and Running via Parallel Composition (2017, background): “Optimization has recently surged in popularity in the literature—with the advent of new tools Tedrake (2009) and the simultaneous boom in computing power—resulting in almostuniversal preference among competitors in the recent DRC (e.g. Kuindersma et al. (2016)).”
From Motor Control to Team Play in Simulated Humanoid Football (2021, background): “Though the efforts by Boston Dynamics are proprietary, they are built from considerable expertise developed by robotics researchers (10).”
The current state and future outlook of rescue robotics (2019, background): “Consequently, many of the robots took on a humanoid morphology (Atkeson et al., 2015; Kohlbrecher et al., 2015; Kuindersma et al., 2016; Feng et al., 2015; Johnson et al., 2015; Tsagarakis et al., 2017; Kaneko et al., 2015).”