- 42
- 72 447
Dynamic Robotics and Control Laboratory
เข้าร่วมเมื่อ 24 มี.ค. 2016
Lab website: sites.usc.edu/quann/
Continuous Dynamic Bipedal Jumping via Adaptive-model Optimization
Preprint: arxiv.org/abs/2404.11807
Title: Continuous Dynamic Bipedal Jumping via Adaptive-model Optimization
Authors: Junheng Li, Omar Kolt, Quan Nguyen
Hardware Platform: HECTOR Humanoid github.com/DRCL-USC/Hector_Simulation
Abstract:
Dynamic and continuous jumping remains an open yet challenging problem in bipedal robot control. The choice of dynamic models in trajectory optimization (TO) problems plays a huge role in trajectory accuracy and computation efficiency, which normally cannot be ensured simultaneously. In this letter, we propose a novel adaptive-model optimization approach, a unified framework of Adaptive-model TO and Adaptive-frequency Model Predictive Control (MPC), to effectively realize continuous and robust jumping on HECTOR bipedal robot. The proposed Adaptive-model TO fuses adaptive-fidelity dynamics modeling of bipedal jumping motion for model fidelity necessities in different jumping phases to ensure trajectory accuracy and computation efficiency. In addition, conventional approaches have unsynchronized sampling frequencies in TO and real-time control, causing the framework to have mismatched modeling resolutions. We adapt MPC sampling frequency based on TO trajectory resolution in different phases for effective trajectory tracking. In hardware experiments, we have demonstrated robust and dynamic jumps covering a distance of up to 40 cm (57% of robot height). To verify the repeatability of this experiment, we run 53 jumping experiments and achieve 90% success rate. In continuous jumps, we demonstrate continuous bipedal jumping with terrain height perturbations (up to 5 cm) and discontinuities (up to 20 cm gap).
(submitted to RA-L)
Title: Continuous Dynamic Bipedal Jumping via Adaptive-model Optimization
Authors: Junheng Li, Omar Kolt, Quan Nguyen
Hardware Platform: HECTOR Humanoid github.com/DRCL-USC/Hector_Simulation
Abstract:
Dynamic and continuous jumping remains an open yet challenging problem in bipedal robot control. The choice of dynamic models in trajectory optimization (TO) problems plays a huge role in trajectory accuracy and computation efficiency, which normally cannot be ensured simultaneously. In this letter, we propose a novel adaptive-model optimization approach, a unified framework of Adaptive-model TO and Adaptive-frequency Model Predictive Control (MPC), to effectively realize continuous and robust jumping on HECTOR bipedal robot. The proposed Adaptive-model TO fuses adaptive-fidelity dynamics modeling of bipedal jumping motion for model fidelity necessities in different jumping phases to ensure trajectory accuracy and computation efficiency. In addition, conventional approaches have unsynchronized sampling frequencies in TO and real-time control, causing the framework to have mismatched modeling resolutions. We adapt MPC sampling frequency based on TO trajectory resolution in different phases for effective trajectory tracking. In hardware experiments, we have demonstrated robust and dynamic jumps covering a distance of up to 40 cm (57% of robot height). To verify the repeatability of this experiment, we run 53 jumping experiments and achieve 90% success rate. In continuous jumps, we demonstrate continuous bipedal jumping with terrain height perturbations (up to 5 cm) and discontinuities (up to 20 cm gap).
(submitted to RA-L)
มุมมอง: 10 283
วีดีโอ
Adaptive Force-based Control of Dynamic Legged Locomotion over Uneven Terrain
มุมมอง 4092 หลายเดือนก่อน
Agile-legged robots have proven to be highly effective in navigating and performing tasks in complex and challenging environments, including disaster zones and industrial settings. However, these applications commonly require the capability of carrying heavy loads while maintaining dynamic motion. Therefore, this paper presents a novel methodology for incorporating adaptive control into a force...
Safety-Aware Perception for Autonomous Collision Avoidance in Dynamic Environments
มุมมอง 1672 หลายเดือนก่อน
Abstract: Autonomous collision avoidance requires accurate environmental perception; however, flight systems often possess limited sensing capabilities with field-of-view (FOV) restrictions. To navigate this challenge, we present a safety-aware approach for online determination of the optimal sensor-pointing direction psi_d which utilizes control barrier functions (CBFs). First, we generate a s...
Learning Agile Locomotion and Agile Behaviors via RL-augmented MPC
มุมมอง 7402 หลายเดือนก่อน
In the context of legged robots, adaptive behavior involves adaptive balancing and adaptive swing foot reflection. While adaptive balancing counteracts perturbations to the robot, adaptive swing foot reflection helps the robot to navigate intricate terrains without foot entrapment. In this paper, we manage to bring both aspects of adaptive behavior to quadruped locomotion by combining RL and MP...
Hierarchical Optimization-based Control for Whole-body Loco-manipulation of Heavy Objects
มุมมอง 3662 หลายเดือนก่อน
In recent years, the field of legged robotics has seen growing interest in enhancing the capabilities of these robots through the integration of articulated robotic arms. However, achieving successful loco-manipulation, especially involving interaction with heavy objects, is far from straightforward, as object manipulation can introduce substantial disturbances that impact the robot's locomotio...
A Fun Day at the Park with HECTOR Robot
มุมมอง 1K2 หลายเดือนก่อน
DRCL team BBQ 2024 at El Dorado Park, Long Beach, ft. HECTOR biped.
Kinodynamic Pose Optimization for Humanoid Loco-manipulation
มุมมอง 8258 หลายเดือนก่อน
This paper presents a novel approach for controlling humanoid robots pushing heavy objects using kinodynamics-based pose optimization and loco-manipulation MPC. The proposed pose optimization plans the optimal pushing pose for the robot while accounting for the unified object-robot dynamics model in steady state, robot kinematic constraints, and object parameters. The approach is combined with ...
DRCL Legged Robot Demonstrations at IROS 2023!
มุมมอง 1.6K8 หลายเดือนก่อน
Thank everyone who has stopped by during our demos! A short video compilation of DRCL legged robot demonstrations at IROS 2023: 1. Dynamic Locomotion on Bipedal Robots Via Force-And-Moment-Based Model Predictive Control on HECTOR (th-cam.com/video/NcW-NFwjMh0/w-d-xo.html) 2. Reinforcement Learning Augmented Model Predictive Control for Quadruped Locomotion
Learning Multimodal Bipedal Locomotion and Implicit Transitions: A Versatile Policy Approach
มุมมอง 3478 หลายเดือนก่อน
IROS 2023 presentation video for paper: "Learning Multimodal Bipedal Locomotion and Implicit Transitions: A Versatile Policy Approach" arXiv: arxiv.org/abs/2303.05711
Introducing Hector - Humanoid for Enhanced ConTrol and Open-source Research
มุมมอง 5K8 หลายเดือนก่อน
Hector will be demonstrated at IROS 2023! Session WeBIP-21.13 on Wednesday, Oct. 4th from 3:30 PM to 5 PM in Exhibition Hall Project Hector combines the knowledge acquired by our team in legged robot control and design in the past years. Our objective is to develop a cost-effective, small-scale, and power-dense humanoid robot hardware platform, which will be accessible to the broader robotics a...
Hierarchical Adaptive Control for Collaborative Manipulation of a Rigid Object by Quadrupedal Robots
มุมมอง 2919 หลายเดือนก่อน
IROS 2023 presentation for the paper "Hierarchical Adaptive Control for Collaborative Manipulation of a Rigid Object by Quadrupedal Robots." Paper link: arxiv.org/abs/2303.06741
Learning Multimodal Bipedal Locomotion and Implicit Transitions: A Versatile Policy Approach
มุมมอง 64510 หลายเดือนก่อน
abstract: In this paper, we propose a novel framework for synthesizing a single multimodal control policy capable of generating diverse behaviors (or modes) and emergent inherent transition maneuvers for bipedal locomotion. In our method, we first learn efficient latent encodings for each behavior by training an autoencoder from a dataset of rough reference motions. These latent encodings are u...
Hierarchical Adaptive Control for Collaborative Manipulation of a Rigid Object by Quadrupedal Robots
มุมมอง 49810 หลายเดือนก่อน
Hierarchical Adaptive Control for Collaborative Manipulation of a Rigid Object by Quadrupedal Robots
Contact Optimization for Non-Prehensile Loco-Manipulation via Hierarchical Model Predictive Control
มุมมอง 229ปีที่แล้ว
arxiv.org/abs/2210.03442
Hierarchical Adaptive Loco-manipulation Control for Quadruped Robots
มุมมอง 301ปีที่แล้ว
arxiv.org/abs/2209.13145
Contact-timing and Trajectory Optimization for 3D Jumping on Quadruped Robots
มุมมอง 680ปีที่แล้ว
Contact-timing and Trajectory Optimization for 3D Jumping on Quadruped Robots
Informed Sampling-Based Planning to Enable Legged Robtos to Safely Negotiate Permeable Obstacles
มุมมอง 294ปีที่แล้ว
Informed Sampling-Based Planning to Enable Legged Robtos to Safely Negotiate Permeable Obstacles
Contact Optimization for Non-Prehensile Loco-Manipulation via Hierarchical Model Predictive Control
มุมมอง 353ปีที่แล้ว
Contact Optimization for Non-Prehensile Loco-Manipulation via Hierarchical Model Predictive Control
Multi-contact MPC for Dynamic Loco-manipulation on Humanoid Robots
มุมมอง 1.4Kปีที่แล้ว
Multi-contact MPC for Dynamic Loco-manipulation on Humanoid Robots
A Hybrid Quadratic Programming Framework for Real-Time Embedded Safety-Critical Control
มุมมอง 221ปีที่แล้ว
A Hybrid Quadratic Programming Framework for Real-Time Embedded Safety-Critical Control
Hierarchical Adaptive Loco-manipulation Control for Quadruped Robots
มุมมอง 395ปีที่แล้ว
Hierarchical Adaptive Loco-manipulation Control for Quadruped Robots
Dynamic Walking of Bipedal Robots on Uneven Stepping Stones via Adaptive-frequency MPC
มุมมอง 886ปีที่แล้ว
Dynamic Walking of Bipedal Robots on Uneven Stepping Stones via Adaptive-frequency MPC
Hierarchical Adaptive Loco-manipulation Control for Quadruped Robots
มุมมอง 634ปีที่แล้ว
Hierarchical Adaptive Loco-manipulation Control for Quadruped Robots
Robust High-Speed Running for Quadruped Robots via Deep Reinforcement Learning
มุมมอง 1.3Kปีที่แล้ว
Robust High-Speed Running for Quadruped Robots via Deep Reinforcement Learning
Continuous Jumping for Legged Robots on Stepping Stones via Trajectory Optimization and MPC
มุมมอง 1.5Kปีที่แล้ว
Continuous Jumping for Legged Robots on Stepping Stones via Trajectory Optimization and MPC
Contact-timing and Trajectory Optimization for 3D Jumping on Quadruped Robots
มุมมอง 1.2Kปีที่แล้ว
Contact-timing and Trajectory Optimization for 3D Jumping on Quadruped Robots
Balancing Control and Pose Optimization for Wheel-legged Robots Navigating High Obstacles
มุมมอง 2.7Kปีที่แล้ว
Balancing Control and Pose Optimization for Wheel-legged Robots Navigating High Obstacles
Integrating Motion Planning and Control for Legged Robots Negotiating Permeable Obstacles
มุมมอง 5952 ปีที่แล้ว
Integrating Motion Planning and Control for Legged Robots Negotiating Permeable Obstacles
CDC 2021 Presentation - Force-and-moment-based MPC for Highly Dynamic Bipedal Robots
มุมมอง 25K2 ปีที่แล้ว
CDC 2021 Presentation - Force-and-moment-based MPC for Highly Dynamic Bipedal Robots
CDC 2021 Presentation - Safe and Robust Motion Planning for Dynamic Robotics via CBFs
มุมมอง 3492 ปีที่แล้ว
CDC 2021 Presentation - Safe and Robust Motion Planning for Dynamic Robotics via CBFs
By the time we use up (and destroy) 96% of our natural resources on Earth, we really should be able to make a robot that has 4% of dexterity and 0.4% of efficiency of the average that nature could, quasi spontaneously produce - and that average already includes plants, fungi and viruses. The few apes who could communicate with humans were so right: people are stupid. Hubris is not a religious, esoteric or archaic expression anymore since we felt it up with exact and tangible meaning.
No one else seeing AT walkers from star wars?
Probably coming sooner than we think, and for similar reasons
what does under-actuation means for bipedal locomotion ?
Whoa - This is amazing!
It is a very good work. Will you release your codes?
Walks slower than older version robots…what’s the point?
has any of the code been made opensource? thank you.
Nice!
real life pixar lamp?
👍👍👍
Isn't this the idea of PE condition for observability? Just that you are choosing nature sine wave as the variable?
HECTOR looks like Footbot from Gravity Falls.
Chicken Walkers
Did you use RL to make it walk?
Very cool! Can I get Hector robot cad file ?
Does it actualky work though , do you have any actual code. Good presentation. Models are cool, matlab is trash, id like to see some code though.
open sourced in both ROS and Matlab: github.com/DRCL-USC/Hector_Simulation
@@drcl_usc checked it out, not that impressed tbh. The code isn't thst grand, it works. It's just incredibly simple and some/ a good bit of its generated code..., ( like you didn't write the architecture ) Good for a presentation in college, not for development tho.
Hi awesome. can I get your thesis if possible ?
it'll be impressive when they can walk and stand like they aren't holding in a sht
大佬厉害，要是能做成 docker 就好了
dope!
None of these are as impressive as the new Disney Robot seen at IROS.
Didn't this guy fall 15 times at that Darpa competition or was that a different Hector? Just kidding, great work
What size shoe does hector wear?
saw ur work in London. how is it different with your ICRA paper?????
oh multi robot case. ok got real robot demo?
More results on robot hardware are coming soon. Keep tuned!
Could the author provide a GitHub link for research purposes?
Excellent presentation. Enjoyed it. Gives me ideas for my own bots!
*Promo SM* ☺️
Nice.
how did you crete this simulations? I am new to this
ETH quadruple robot is strong. but USC robot is fun
lol very interesting
Your work is really interesting. Keep going.
Dear Mohsen you are the best.
Is the paper not published yet?
How to keep stable while walking?
ƤＲＯ𝓂Ｏ𝕤ᗰ 💥
Awesome! Have you tried to deploy this on a physical system?
wow nice so underrated
That’s a really cool project. I wonder how did you modify the existing Unitree Robotics hardware to have actuated wheels. Are there such components available to buy?
Nicely executed 👏👏
Great work ! Does this implementation run in real time ?
Very nice and useful video… Thank you... Please LIKE and SUBSCRIBE next: th-cam.com/video/S1zEeRmHfrc/w-d-xo.html th-cam.com/video/6OTDTO_cOXI/w-d-xo.html th-cam.com/video/VEZ-eqUInAg/w-d-xo.html
Hardware results?
Nice work! Do you make the code public?