C's Robotics Paper Notes
@RoboReading
Followers
3K
Following
10
Media
266
Statuses
459
Recording daily paper notes by @chongzitazhang who is interested in robot learning and control, with special focus on legged robots. ///I'm picky///
Joined March 2024
https://t.co/9YCkScsvFC KiVi: Kinesthetic-Visuospatial Integration for Dynamic and Safe Egocentric Legged Locomotion 1. learning encoder-decoder things separately for proprioception and exteroception 2. better memory arch 3. noise randomization enables robust visual locomotion
1
5
21
Towards Versatile Humanoid Table Tennis: Unified Reinforcement Learning with Prediction Augmentation humanoid table tennis with reward engineering and concurrent prediction learning
1
4
10
Push Anything: Single- and Multi-Object Pushing From First Sight with Contact-Implicit MPC https://t.co/2RSd92cGpq push diverse objects via sampling based CIMPC
0
1
6
Reinforcement Learning-based Robust Wall Climbing Locomotion Controller in Ferromagnetic Environment https://t.co/Ssjk9AUej7 curriculum learning + foot adhesion modelling
0
1
12
The Reality Gap in Robotics: Challenges, Solutions, and Best Practices https://t.co/K7arEyIKwu Reading the "Symptoms" reminds me of many bad memories 😅
1
8
56
VAMOS: A Hierarchical Vision-Language-Action Model for Capability-Modulated and Steerable Navigation 1) VLM planner proposes traj candidates 2) affordance model filters trajs.
1
7
71
softmimic: learning compliant wbc from examples https://t.co/nqfoyVPeLN 1. stiffness based reward 2. augmentation of references compliance enables a bunch of downstream applications such as loco-manipulation, and also helps robustness.
0
0
1
https://t.co/BvDzsvgd4k Humanoid Goalkeeper: Learning from Position Conditioned Task-Motion Constraints multi ref amp + task reward design
0
0
12
https://t.co/1eYCywW6I5 SEMANTIC WORLD MODELS world models going beyond pixels and do semantic reasoning for the future
0
2
10
HDMI: Learning Interactive Humanoid Whole-Body Control from Human Videos https://t.co/yD6vkd6ESu tracking both human and object states, use ref in actions, use contact rewards and ref.
0
3
14
ResMimic: From General Motion Tracking to Humanoid Whole-body Loco-Manipulation via Residual Learning https://t.co/76kcynpThj GMT + residual policy for stable interactive loco-mani wbc
0
11
47
https://t.co/zNfH3xjK0g Humanoid Everyday: A Comprehensive Robotic Dataset for Open-World Humanoid Manipulation
0
1
10
https://t.co/6BNZ3uYwty COLA: Learning Human-Humanoid Coordination for Collaborative Object Carrying base tracking policy + residual collaboration teacher policy => student collaborative wbc
0
0
0
https://t.co/sV1IEOEIAr adamimic residual phase change + residual action for repurposing tracking policies.
0
0
3
Robot Learning from Any Images https://t.co/kvK1z9gOgV Generating physical robotic environments from images
0
1
11
LocoFormer: Generalist Locomotion via Long-context Adaptation https://t.co/XdULwAmKEN RL for diverse morph at scale with cross-episode memory to enable long-context adaptation
1
2
25