Robotics & Embodied AI - Weekly Roundup

Interactive 3D asset generation can now be driven by functional logic and hierarchical physics, thanks to a new framework that synthesizes simulation-ready assets.

Stop committing to a single policy in offline-to-online RL: adaptively select and fine-tune policies based on predicted performance to maximize returns under interaction budgets.

Stabilizing nuclear fusion plasma with imitation learning is possible even with limited macroscopic observations, offering a path to practical control strategies.

Drivers dynamically switch their perceptual priorities from gap-closing rate to visual looming as braking intensity decreases, overturning long-held assumptions about car-following behavior.

Task-aware 3D reconstruction slashes the number of views needed by focusing on the data that actually matters for downstream applications.

Interactive 3D asset generation can now be driven by functional logic and hierarchical physics, thanks to a new framework that synthesizes simulation-ready assets.

Stop committing to a single policy in offline-to-online RL: adaptively select and fine-tune policies based on predicted performance to maximize returns under interaction budgets.

Stabilizing nuclear fusion plasma with imitation learning is possible even with limited macroscopic observations, offering a path to practical control strategies.

Drivers dynamically switch their perceptual priorities from gap-closing rate to visual looming as braking intensity decreases, overturning long-held assumptions about car-following behavior.

Task-aware 3D reconstruction slashes the number of views needed by focusing on the data that actually matters for downstream applications.

You don't need a full causal graph to avoid undesired outcomes; learning a simple order structure can be enough, and even outperform methods that try to learn the whole graph.

Reservoir Computing offers a surprisingly effective way to build Koopman dictionaries for nonlinear system identification, sidestepping the usual dictionary selection and ill-conditioning problems.

Average reward RL can finally handle the messy reality of non-stationary rewards and durations in SMDPs, thanks to a clever harmonic mean trick.

Control-dependent latent dynamics, achieved with a surprisingly small parameter increase, unlock robust MPC performance in time-varying environments where standard Koopman methods falter.

Turns out, all gaze estimation models stumble when robots look down, and complex architectures aren't the answer – data diversity is the real secret to robust human-robot interaction.

MoEs, despite their scaling advantages, suffer from a surprising "spectral plasticity loss" in continual RL, but a simple Parseval penalty can recover performance.

Achieve robust long-horizon visual control by adaptively balancing model-based lookahead with bootstrapping, enabling zero-shot transfer to real-world tasks with severe occlusions.

Decomposing robot swarm state representations unlocks effective cooperation even with computationally-limited agents.

Forget brittle imitation learning: Q2RL unlocks robust on-robot reinforcement learning by distilling a Q-function from Behavior Cloning and intelligently gating between imitation and RL based on Q-value estimates.

Forget hand-crafted reward functions: this RL framework lets a bicycle robot learn complex stunts from just a spatial guideline and a few key poses.

Predicting driver behavior in response to traffic conditions is now possible with a new world model that causally links external context to internal driver states.

End-to-end ML models get smoked in real-world mmWave vehicular connectivity: a hybrid vision-primed approach slashes outage rates by leveraging model-based reasoning and RF feedback.

Fragmented privacy patches are insufficient for Embodied AI: a unified, lifecycle-level approach is needed to prevent systemic privacy leakage in real-world deployments.

LLM-powered multi-agent collaboration can boost zero-shot IMU activity recognition accuracy by 29% compared to existing agent models, even surpassing deep learning baselines.

Gradient-based MPC can finally beat gradient-free methods in continuous control, thanks to Dream-MPC's clever combination of learned policies, world models, uncertainty regularization, and optimization amortization.

DAOs could unlock a new era of human-machine collaboration by democratizing the operation and governance of physical-digital systems.

Unlock zero-shot 3D scene understanding: Ilov3Splat lets you identify and segment arbitrary objects in 3D scenes using only natural language, no category supervision needed.

Mixing tasks with different safety levels in automotive ECUs can compromise critical functions, highlighting the need for careful task allocation strategies.

Guaranteeing safety in autonomous systems gets a boost: this work enables formal verification of hybrid system code that directly controls physical processes.

Get high-fidelity tactile simulations with 65% speedup and 40% less memory by combining coarse physics with neural implicit reconstruction.

By intelligently incorporating LiDAR-derived height information, HiPR overcomes limitations of fixed projection spaces, achieving state-of-the-art camera-LiDAR occupancy prediction with real-time performance.

Finally, a driving dataset that doesn't just assume perfectly paved roads, offering 6.5x more depth data than KITTI for realistic autonomous driving scenarios.

Adult-trained human mesh recovery models can now handle kids, too, thanks to a new framework that enforces spatial consistency and leverages VLM-derived age and gender cues.

Synthesizing realistic duet dance motions gets a boost from explicitly modeling inter-dancer contact, leading to significantly improved interaction fidelity and rhythmic synchronization.

Bridging the gap between aerial and ground-level tracking, VL-UniTrack uses visual-language prompts to achieve robust object tracking even with significant viewpoint differences.

Radar SLAM can now achieve state-of-the-art performance via direct scan registration, eliminating the need for hand-engineered feature extraction and enabling robust localization in adverse weather.

Autonomous driving gets a boost: CRAFT cleverly combines the best of both worlds – dense counterfactual supervision and grounded closed-loop feedback – to significantly improve driving policies.

Robots can reliably hand over objects to humans by actively probing grasps, achieving a 30% improvement over passive methods.

RL fine-tuning unlocks a 6x performance gain for in-place trajectory editing in autonomous driving, demonstrating the power of aligning diffusion planners with reinforcement learning.

Stop relying on significance tests that only find differences: this Bayesian framework tells you if your synthetic data is *practically equivalent* to real-world data for your specific safety assessment task.

Tactile feedback, when strategically sampled and evaluated, unlocks robust and safe robotic insertion policies even under sub-millimeter tolerances.

Achieve real-time bipedal walking control by cleverly swapping high-fidelity for low-fidelity models in MPC, slashing computation without sacrificing stability.

Diffusion models can now plan effectively for long-horizon tasks by strategically generating subgoals that are then efficiently realized by rectified flow models.

Generate more realistic and diverse safety-critical autonomous vehicle scenarios by using conditional latent flow matching to bridge the gap between real-world and simulated data.

Dynamically adjusting trajectory optimization based on real-time navigation confidence enables robust low-thrust rendezvous, slashing miss distances by two orders of magnitude when faced with degraded sensor data.

Achieve autonomous laparoscope control by translating multimodal surgical data into a single "wrench" that guides the robot's movements.

Forget digital watermarks – now you can physically fingerprint solutions with electrochemically-generated polymer patterns, opening doors to low-cost, physically-encrypted personal information.

Hand-eye calibration gets a 67% accuracy boost in high-uncertainty scenarios thanks to a new optimization framework that cleverly avoids explicit uncertainty modeling.

AI is enabling a new generation of AUV navigation systems that overcome the limitations of traditional model-based approaches in complex underwater environments.

Forget complex assembly: this 3D printing technique lets you pop out functional, self-folding robots with integrated sensors and actuators directly from a flat sheet.

Robotic manipulation gets a serious upgrade: ConsisVLA-4D boosts performance by up to 41.5% and speeds up inference by 2.4x, all while ensuring your robot understands the scene in 3D *and* how it changes over time.

Guaranteeing robot safety in unknown environments doesn't require complex planning – this closed-form CBF filter does it with minimal computation.

Standard camera auto-exposure is blind to the needs of remote heart-rate monitoring, but a new method closes the gap to enable robust in-vehicle driver monitoring.

By grounding temporal Gaussian aggregation in spatial voxels, Ground4D achieves state-of-the-art 4D reconstruction in challenging off-road environments where existing methods falter.

Stop feeding LLMs redundant and conflicting sensor data in autonomous driving: a new architecture slashes hallucinated entities by coordinating multi-sensor inputs *before* reasoning.

Even with noisy initial matches, Angle-I2P leverages angular consistency and hierarchical attention to achieve state-of-the-art image-to-point cloud registration.

Explicitly modeling human-object interactions boosts multi-person human mesh recovery accuracy by up to 9.9%, showing that interaction context is key to understanding human pose and shape in complex scenes.

Image-based latent actions are your secret weapon for long-horizon reasoning in VLAs, while action-based latent actions unlock complex motor coordination.

Alpha-blending, a core optimization in 3D Gaussian Splatting, subtly hobbles feature learning, but a geometry-weighted fusion approach can unlock more accurate and efficient visual localization.

Top-view RGB-D person re-identification is surprisingly feasible, even across modalities, despite the inherent challenges of viewpoint and modality variations.

RLDX-1 achieves double the success rate of existing VLAs on complex humanoid tasks, suggesting a leap in robots' ability to handle contact-rich, dynamic manipulation.

Current world models struggle with basic physical interaction tasks like distance perception and trajectory following, highlighting a critical gap in their ability to simulate realistic environments.

Instead of creating new AI companions from scratch, Deco shows how to breathe new life into cherished physical objects by giving them a digital voice and personality powered by LLMs.

Immersive video reveals that "being there" hinges more on feeling spatially located than having a virtual body, challenging conventional notions of embodiment in XR.

Event cameras can significantly boost the reliability of autonomous driving in high-dynamic-range and high-speed scenarios, achieving perfect route completion in CARLA benchmarks.

Fixed confidence thresholds are holding back explainable autonomous driving systems, but this new adaptive approach and dataset could unlock better performance and cross-cultural understanding.

Automating stage lighting control across diverse venues is now possible without expert demonstrations, thanks to a novel imitation learning approach that decomposes global color distributions into individual light controls.

An open-source alternative to expensive, proprietary digital human modeling software could democratize ergonomic analysis and workplace design.

Guaranteeing safe robot navigation in unstructured environments just got easier: translate human language rules into formal logic, ground them with VLMs, and let the robot navigate.

Achieve a 2.9x reduction in end-to-end latency in ROS 2 communication by trading off scalability for simplicity in cross-process object lifetime management.

End-to-end learning can beat even the best industrial solvers at multi-agent task assignment, improving solution quality by 20% while slashing computation time from hours to seconds.

Quadrupedal robots can now perform dynamic loco-manipulation in the real world, matching human teleoperation, using only onboard ego-centric vision and a low-frequency (5Hz) open-vocabulary detector.

Ditching load sensors and directly embedding cable constraints into the quadrotor's control loop unlocks more precise and robust aerial manipulation.

Guaranteeing safety and liveness in complex control systems doesn't require monolithic design; this work shows how to decompose the problem across layers with formal contracts.

Forget tedious manual tuning: ScanHD lets robots autonomously configure laser profilers based on natural language instructions and visual context, achieving >92% accuracy in real-world inspection tasks.

LLMs alone can't reliably fly drone swarms from natural language commands; task-specific tools and runtime guardrails are essential for real-world cyber-physical system control.

Reactive dexterous grasping can be achieved with zero-shot transfer to real-world objects by decoupling high-level RL planning from low-level QP control, enabling dynamic adjustments to safety margins without retraining.

LLMs spontaneously exhibit collaborative behaviors like perspective-taking and theory of mind in embodied settings, suggesting a surprising capacity for modeling human collaborators without explicit training.

Achieve 15% faster order completion in warehouse robotics with a new deep reinforcement learning approach that jointly optimizes robot scheduling and order allocation in real-time.

Robot video world models can be significantly improved by distilling a multimodal reward function and stabilizing long-horizon inference, leading to better instruction following and manipulation accuracy.

Classical SLAM algorithms crumble under visual degradation, but deep learning approaches like MASt3R and DUSt3R maintain impressive localization accuracy, suggesting a path to robust UAV autonomy in challenging environments.

Achieve near order-of-magnitude reduction in tail timing error in mixed-criticality robotics by decoupling safety-critical control from user applications.

Robots can now learn manipulation skills from human videos with greater morphological accuracy and temporal consistency, thanks to a new method that disentangles task and embodiment.

Achieve scalable open-vocabulary semantic maps of entire buildings by fusing both dense and instance-level semantic information in a novel dual-layer voxel representation.

Escape deadlocks and choreograph robots through complex tasks with this new hybrid control architecture that merges planning and control.

Unlock agile humanoid robots by ditching teleoperation and training directly from human VR demos.

Ditch the GPS: This CVGL pipeline achieves a 5.9x improvement in localization accuracy over IMU-only by intelligently fusing satellite imagery with inertial measurements, needing only a single initial GPS fix.

Domain randomization doesn't just make your robot policies more robust; it fundamentally warps the optimization landscape, potentially guiding your search towards better contact-rich behaviors.

Differentiating through physical simulations just got a whole lot easier: Neural Control avoids unrolling iterative solvers by using an adjoint formulation, enabling memory-efficient gradient-based control.

Autonomous vehicles can now stick to the plan even with disturbances, thanks to a new control method that learns and compensates for unmodeled dynamics.

Control heterogeneous physical neural networks—even wetware—with a single orchestration architecture, opening the door to seamless integration with edge-cloud workflows.

Macromolecules surf differently: they exhibit mixed frictional behavior on hydrophobic surfaces, rubbing against the solid while being dragged by the flow, unlike their purely advective transport on hydrophilic surfaces.

Open-sourcing a VLA model that beats closed-source giants on embodied reasoning tasks could finally make real-world robot deployment practical.

Rapidly prototype sensor-driven applications across diverse infrastructures without needing cross-domain expertise using AI-assisted, pattern-based workflow engineering.

CAVs can now detect sensor anomalies in their measurements without relying on a central unit, even when tracking human-driven vehicles that aren't directly observable.

Combining diffusion models with image-to-image translation yields surprisingly realistic synthetic data, outperforming either method alone in closing the sim2real gap.

Current MLLM-driven UAV agents still struggle with spatial memory and aerial adaptation when tasked with autonomously exploring and reasoning about victim locations in realistic search and rescue scenarios.

LLMs can now intelligently orchestrate multi-agent systems, learning to optimize both individual agent actions and inter-agent cooperation for distributed black-box problems.

Generalist robot policies can achieve 95% success rates on real-world manipulation tasks by continually learning from a fleet of robots, even in the face of distribution shifts and long-tail failures.