Robotics
Research on robot control, manipulation, navigation, and human-robot interaction.
cs.RO · 524 papersAccurate Trajectory Tracking with MPCC for Flapping-Wing MAVs
A novel MPCC method enables accurate, real-time trajectory tracking for flapping-wing MAVs, significantly improving control precision.
A Closed-Form Dual-Barrier CBF Safety Filter for Holonomic Robots on Incrementally Built Occupancy Grid Maps
This paper introduces a dual-barrier CBF safety filter for holonomic robots, ensuring real-time collision avoidance in unknown environments using occupancy grid maps.
When Life Gives You BC, Make Q-functions: Extracting Q-values from Behavior Cloning for On-Robot Reinforcement Learning
Q2RL extracts Q-values from Behavior Cloning to enable efficient on-robot reinforcement learning, improving policies online without losing learned skills.
ConsisVLA-4D: Advancing Spatiotemporal Consistency in Efficient 3D-Perception and 4D-Reasoning for Robotic Manipulation
ConsisVLA-4D enhances robotic manipulation by improving spatiotemporal consistency in 3D perception and 4D reasoning, achieving significant speedups.
LineRides: Line-Guided Reinforcement Learning for Bicycle Robot Stunts
LineRides is a novel line-guided reinforcement learning framework enabling a bicycle robot to learn diverse stunts without demonstrations.
Driver-WM: A Driver-Centric Traffic-Conditioned Latent World Model for In-Cabin Dynamics Rollout
Driver-WM is a novel latent world model forecasting in-cabin driver dynamics, causally conditioned on external traffic, for safer L2/L3 autonomous driving.
Reduced-order Neural Modeling with Differentiable Simulation for High-Detail Tactile Perception
A reduced-order neural simulation framework enables high-detail, physically grounded tactile perception with significant efficiency gains for robotics.
Position: Embodied AI Requires a Privacy-Utility Trade-off
Embodied AI systems face a systemic privacy crisis; this paper argues for privacy as a lifecycle architectural constraint, proposing the SPINE framework.
Modular Reinforcement Learning For Cooperative Swarms
This paper introduces a modular reinforcement learning approach to efficiently represent spatial interaction states in cooperative robot swarms, overcoming memory challenges.
Koopman Identification of Nonlinear Systems via Reservoir Liftings
RC-Koopman uses reservoir computing to create a well-conditioned, stateful dictionary for Koopman operator identification of nonlinear systems.
Optimal Uncertainty-Aware Calibration for the AX=YB Problem
This paper introduces an uncertainty-aware Lie algebra-based optimization for AX=YB hand-eye calibration, achieving global optimal solutions and significantly improving accuracy.
Dr-PoGO: Direct Radar Pose-Graph Optimization
Dr-PoGO is a novel radar-based SLAM method using direct registration and pose-graph optimization for robust localization in all weather conditions.
Gaze4HRI: Zero-shot Benchmarking Gaze Estimation Neural-Networks for Human-Robot Interaction
Gaze4HRI introduces a large-scale benchmark for zero-shot gaze estimation in HRI, revealing current methods' failures and highlighting data diversity as key to robustness.
3D Printing of Passively Actuated Self-Folding Robots with Integrated Functional Modules
A 3D printing method creates passively actuated, self-folding robots from conductive PLA nets, integrating electronics and a predictive folding model.
ELVIS: Ensemble-Calibrated Latent Imagination for Long-Horizon Visual MPC
ELVIS enables robust long-horizon visual MPC via Gaussian-mixture planning and an uncertainty-aware lambda-return, handling branching futures and model errors.
From Pixels to Tokens: A Systematic Study of Latent Action Supervision for Vision-Language-Action Models
This paper systematically compares latent action supervision methods for VLA models, finding image-based actions aid reasoning and action-based actions improve motor skills.
AI-Aided Advancements in Autonomous Underwater Vehicle Navigation
This paper reviews AI-aided advancements in AUV navigation, focusing on sensor fusion architectures and AI-driven learning for high-precision positioning.
From Reach to Insert: Tactile-Augmented Precision Assembly under Sub-Millimeter Tolerances
A tactile-augmented two-stage method combining IL and RL achieves high-precision robotic insertions with sub-millimeter tolerances and low forces.
ReflectDrive-2: Reinforcement-Learning-Aligned Self-Editing for Discrete Diffusion Driving
ReflectDrive-2 introduces a discrete diffusion planner for autonomous driving with self-editing capabilities, significantly improved by reinforcement learning.
Active Contact Sensing for Robust Robot-to-Human Object Handover
Robots use active contact sensing and information-gathering motions to robustly hand over objects, achieving 97.5% success.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.