archive
Every paper Pith has read. Search by title, abstract, or pith.
4720 papers in cs.RO · page 1
-
Tactile deformation prediction lifts contact task success to 71.67%
VT-WAM: Visual-Tactile World Action Model for Contact-Rich Manipulation
-
C++ runtime runs embodied models on any robot hardware
Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots
-
Behavior latents add independent speed and safety control to traffic sims
Controllable Sim Agents with Behavior Latents
-
QuadRocket achieves almost global trajectory tracking with adaptive control
QuadRocket: An Aerial Robotic Testbed for Adaptive Thrust-Vector Control of Rocket-Like Vehicles
-
Quadrotor learns to intercept using only camera direction vectors
Learning Agile Intruder Interception using Differentiable Quadrotor Dynamics
-
Unlabeled robot play pretraining matches million-expert VLA models
Learning to Move Before Learning to Do: Task-Agnostic pretraining for VLAs
-
Robot RL success rises 28% with closed world-model loop
WorldSample: Closed-loop Real-robot RL with World Modelling
-
Model learns intent-driven camera poses from passive video
LIME: Learning Intent-aware Camera Motion from Egocentric Video
-
Inverse-dynamics check cuts planning compute in world models
ACID: Action Consistency via Inverse Dynamics for Planning with World Models
-
Humanoid carries 24kg while tracking VR commands
HEFT: Heavy-Payload Full-size Humanoid Teleoperation with Privileged Motion Guidance and Windowed Payload Curriculum
-
Hybrid moving-static camera data cuts VLA shortcut learning
The Moving Eye: Enhancing VLA Spatial Generalization via Hybrid Dynamic Data Collection
-
Low-cost drone runs real-time face tracking and line following
Real-Time Visual Intelligence on Low-Cost UAVs: A Modular Approach for Tracking, Scanning, and Navigation
-
Neuro-symbolic UAV succeeds in 61 of 72 landing tests
NEUROSYMLAND: Neuro-Symbolic Landing-Site Assessment for Robust and Edge-Deployable UAV Autonomy
-
Flow fields outperform action baselines in continuous robot navigation
CoFL-S: Spatially Queryable Sector Flow Fields for Local Language-Conditioned Navigation
-
Shaping real actuators to sim model enables zero-shot policy transfer
Actuator Reality Shaping for Zero-Shot Sim-to-Real Robot Learning
-
Change priors replace full future images for robot actions
Bridge-WA: Predicting Where and How the World Changes for Robotic Action
-
Browser studio turns robot boats into music choreographers
Choreographing the Way of Water: A Computational Framework for Aquatic Robotic Art
-
RBF activation function shapes robotic tracking performance
Influence of Radial Basis Activation Functions on Intelligent Controller for Robotic Manipulators
-
Critic guidance raises frozen VLA success from 68% to 82%
Guided Action Flow: Q-Guided Inference for Flow-Matching Vision-Language-Action Policies
-
One RL policy tracks trajectories on two different boats after simulation training
Cross-Platform Control for Autonomous Surface Vehicles via Adaptive Reinforcement Learning
-
Cross disparity flags moving points for accurate SLAM
A Stereo Visual SLAM System Using Object-Level Motion Estimation and Geometric Filtering Based on Cross Disparity
-
Consolidation updates knowledge while identity hash stays byte-equal
Episodic-to-Semantic Consolidation Without Identity Drift
-
Manifold projections extract novel views from any video model
NeoMap: Training-free Novel-View Synthesis from Single Images and Videos
-
Divergence-free 3D Gaussian field improves robot dynamic manipulation
PhysMani: Physics-principled 3D World Model for Dynamic Object Manipulation
-
Social preference data replaces manual rewards for crowd robot navigation
SPLC: Social Preference Learning for Crowd Robot Navigation
-
Staged scattering model narrows underwater vision gap
Robust Image Processing Techniques for Construction Environment Monitoring Using Underwater Robots
-
Object-level probabilities prune dynamic Gaussians for cleaner SLAM
DL-SLAM: Enabling High-Fidelity Gaussian Splatting SLAM in Dynamic Environments based on Dual-Level Probability
-
VLA-Corrector turns fixed action chunks into on-demand adaptive horizons
VLA-Corrector: Lightweight Detect-and-Correct Inference for Adaptive Action Horizon
-
Pixel diffusion creates 3D Gaussian splats directly in one pass
PixGS: Pixel-Space Diffusion for Direct 3D Gaussian Splat Generation
-
Lightweight safe RL navigates UAVs through dense obstacles
Lightweight Safe Reinforcement Learning for End-to-End UAV Navigation
-
Learned front-ends cut VI-SLAM error by up to 38 percent on some datasets
DL-VINS-Factory: A Modular Framework for Learned Visual Front-Ends in Visual-Inertial SLAM
-
Vision models build hybrid rewards for aligned robot policies
CoRe: Combined Rewards with Vision-Language Model Feedback for Preference-Aligned Reinforcement Learning
-
Imagined tactile signals boost robot tasks up to 44 percent
Imagining the Sense of Touch: Touch-Informed Manipulation via Imagined Tactile Representations
-
One demonstration automates real-world robot RL
One Demonstration Is Enough for Real-World Robotic Reinforcement Learning
-
USV local planner routes behind moving obstacles
Path planning for unmanned naval surface vehicles
-
Language and future signals stabilize VLA transfer on mixed robot data
VLAFlow: A Unified Training Framework for Vision-Language-Action Models via Co-training and Future Latent Alignment
-
MR-NMPC raises quadruped success rate 2.9x on rough terrain
Multi-Rate Nonlinear Model Predictive Control for Wall-Supported Bipedal Locomotion of Quadrupedal Robots
-
Robot turns in place five times faster by dropping to four wheels
A Reconfigurable Rocker-Bogie Robot for High Step Climbing and Turning
-
Scene text forces LVLMs to overthink and slows robots up to 7x
Overthink-Triggered Slowdown Attacks on LVLM-Based Robotic Systems
-
SE(2) mesh captures over 50 percent more traversable area
SE(2) Navigation Mesh
-
BIFROST creates shared latents for zero-shot sim-to-real robot transfer
BIFROST: Bridging Invariant Feature Representation for Observation-space Sim2Real Transfer
-
Simulator syncs time to wall clock for live human-AV planner tests
CommonRoad-Game: A Human-in-the-Loop Simulation Framework for Autonomous Driving
-
Flow matching safety guidance reaches 82.8% collision avoidance
Neuro-Symbolic Safety Guidance for Vision-Language-Action Models via Constrained Flow Matching
-
Generalized reward gives orbit agents control over image timing
Simulation Based Reward Function Validation for Multi-Agent On Orbit Inspection
-
Progress signal raises bimanual assembly success to 80%
FurnitureVLA: Learning Long-Horizon Bimanual Furniture Assembly with Vision-Language-Action Model
-
The paper develops GPU-parallel methods to compute tight linearization error bounds for…
GPU-Parallel Linearization Error Bounds for Real-Time Robust Optimal Control of Nonlinear and Neural Network Dynamics
-
Inverse dynamics model replaces force sensors in four-channel teleop
Sensorless Four-Channel Control Architecture Using Inverse Dynamics Modeling for Human-Scale Bilateral Teleoperation
-
Quadrotor safety filter on 3DGS cuts jerk 47%
FastBridge: Closing the Model-Based Realization Gap in Safety Filters on 3D Gaussian Splatting for Fast Quadrotor Flight
-
4D latent model improves robot planning with 3D consistency
Structured 4D Latent Predictive Model for Robot Planning
-
Metric-free training tops every Waymo forecasting score
Towards Metric-Agnostic Trajectory Forecasting