Robotics & Embodied AI - Weekly Roundup

Achieve high-precision multi-robot SLAM with minimal data transmission by selectively compressing and transmitting keyframes and non-keyframes in a cloud-edge-robot architecture.

Unlock zero-shot sim-to-real transfer for complex legged robots by offloading gait selection to a learned policy that guides a lower-level MPC.

Tactile robotic perception gets a boost with a new pretraining method that explicitly encodes force, geometry, and orientation, leading to a 52% reduction in regression error.

AssistMimic enables humanoid robots to learn complex, force-exchanging assistive motions by reformulating imitation learning as a multi-agent RL problem.

Ditch the pre-trained models: TacLoc achieves accurate robotic pose estimation from tactile sensing alone by framing it as a one-shot point cloud registration problem.

Unlock zero-shot sim-to-real transfer for complex legged robots by offloading gait selection to a learned policy that guides a lower-level MPC.

Tactile robotic perception gets a boost with a new pretraining method that explicitly encodes force, geometry, and orientation, leading to a 52% reduction in regression error.

AssistMimic enables humanoid robots to learn complex, force-exchanging assistive motions by reformulating imitation learning as a multi-agent RL problem.

Ditch the pre-trained models: TacLoc achieves accurate robotic pose estimation from tactile sensing alone by framing it as a one-shot point cloud registration problem.

Achieve up to 1.28x faster VLA model inference for robotic manipulation without retraining, simply by merging visual tokens based on depth.

Multi-robot coverage can now handle multiple sensory demands simultaneously, with provable guarantees on performance even when those demands are initially unknown.

Self-supervised learning can dramatically improve online HD map construction, outperforming supervised methods even with limited labeled data by leveraging geospatial consistency in BEV feature representations.

Achieve real-time safety-critical robot control in partially observable environments by decoupling goal reaching, information gathering, and safety into modular, certificate-based components operating directly in belief space.

VLA-controlled robots can now detect anomalies in under 100ms using a plug-and-play module, enabling real-time recovery from unexpected situations.

Stop wrestling with unstable action spaces: ResWM reframes visual RL by predicting incremental action adjustments, leading to smoother control and better performance.

Unlock bimanual-level cloth manipulation with a single robotic arm using a novel tactile gripper and vision-based perception framework.

Robots can boost their perceived competence by 83% simply by tweaking navigation behaviors suggested by a causal Bayesian network.

Autonomous driving's next leap hinges on reasoning, not just perception, but current LLM-based approaches are too slow for real-time control.

Ditch the clunky controllers: this hand-shadowing pipeline lets you teleoperate a robot arm with just an RGB-D camera and some clever inverse kinematics.

Hyper-redundant robots get a 75% accuracy boost thanks to a neural network that adaptively blends learned behavior with kinematic priors.

Humanoid robots can now reliably transport objects on a tray in the real world, thanks to a hierarchical RL approach that isolates and cancels gait-induced disturbances.

Forget hand-tuning rollout budgets: $V_{0.5}$ dynamically allocates compute to sparse RL rollouts based on a real-time statistical test of a generalist value model's prior, slashing variance and boosting performance.

LVLMs can now provide depth-aware pedestrian navigation guidance by grounding language reasoning and segmentation, without needing user-provided cues or anchor points.

Achieve the seemingly impossible: ASTER uses RL to enable cable-suspended quadrotors to perform autonomous inverted flight.

Robots lost in the vineyard? Not anymore: encoding row-level semantics into a particle filter enables robust localization in repetitive agricultural environments where LiDAR and vision alone fail.

Guarantee runtime safety in complex cyber-physical systems with unbounded data domains using a refinement type system for parameterized streams, even though it's generally undecidable.

Forget training on massive datasets: this new diffusion policy learns human-like 3D scanning strategies that generalize to unseen objects while being robust to noise.

Training embodied intelligence models just got 40x faster thanks to a thousand-GPU cloud platform and a suite of optimizations spanning data pipelines, model architecture, and infrastructure.

Forget catastrophic forgetting: this imitation learning framework remembers up to 65% more while improving AUC by 10-17 points on the LIBERO benchmark.

Finally, a multi-robot path planning benchmark that lets you directly compare grid-based, roadmap, and continuous planners on the same tasks.

Even in feature-rich environments, LiDAR SLAM systems are vulnerable to a new spoofing attack (D-SLAMSpoof) that injects dynamically coordinated spurious point clouds, but can be defended against using inertial dead reckoning.

Steer your robot's diffusion policy away from failure modes at inference time with a lightweight performance predictor trained via self-supervised attention.

Forget hand-crafted rewards: this new method learns dexterous manipulation by encouraging the robot hand to explore diverse contact patterns on objects, leading to impressive real-world transfer.

Robust co-design optimization can significantly improve the performance of agile UAVs in real-world environments by directly incorporating uncertainty and disturbances into the design process.

Robots can now adaptively decide whether to clear clutter or directly grasp, leading to significantly improved success rates in densely cluttered environments.

Forget signal injection – a strategically placed, actuated mirror can now hijack even the most secure LiDAR SLAM systems, inducing localization errors exceeding 6 meters.

Achieve robust humanoid task execution in complex environments by turning high-level language instructions into verifiable, geometrically-grounded task programs that can recover from failures.

By forecasting compact world dynamics before taking action, DynVLA leapfrogs traditional CoT methods to achieve more informed and physically grounded autonomous driving decisions.

By adaptively weighting neighbor information based on uncertainty, distributed multi-object tracking can achieve significantly better performance in mobile robot networks with heterogeneous localization quality.

Multi-robot systems can slash battery consumption by 15% and boost GPU utilization by 50% for large DNN inference by using a hybrid offline-online reinforcement learning strategy to dynamically schedule and distribute DNN module execution.

Robots can now learn to manipulate novel objects in dynamic environments by using LLMs to bridge the gap between symbolic planning and reinforcement learning.

Unlock superior trajectories in complex environments with a new ADMM-based solver that jointly optimizes spatial and temporal domains, eliminating the need for complex warm starting.

Forget sequential robot moves: coordinated "amoebot" swarms can morph into target shapes in near-instant time.

Incomplete trajectory data got you down? This plug-and-play framework progressively aligns features from incomplete observations with complete ones, boosting prediction accuracy in autonomous driving scenarios.

Achieve 2.5x higher success in UAV navigation by decoupling target generation from progress monitoring, enabling safer and more efficient zero-shot flight.

Trajectory optimization just got a whole lot faster and more energy-efficient: a GPU-native solver achieves 4x speedup and halves energy consumption compared to optimized CPU baselines.

Achieve 2x better coverage of autonomous driving safety requirements with 6x fewer simulations by automatically generating test scenarios from formal LTLf specifications.

Injecting muscle synergy priors into reinforcement learning drastically improves the realism of simulated human locomotion, even with limited real-world data.

Reaction wheels can dramatically stabilize bipedal hopping robots in low-gravity environments, enabling more consistent upright landings on irregular extraterrestrial terrains.

Achieve significantly higher accuracy and lower mental demand in bimanual teleoperation by intelligently coupling intention estimation with scene-graph task planning and context-aware motion assistance.

A quadruped robot can now autonomously navigate rough terrain and pick up trash, potentially revolutionizing environmental cleanup in areas inaccessible to traditional robots.

Monocular depth estimation can now run at 161 FPS on edge devices without sacrificing too much accuracy, thanks to a clever asynchronous architecture that reuses features from a foundation model.

Achieve high-precision multi-robot SLAM with minimal data transmission by selectively compressing and transmitting keyframes and non-keyframes in a cloud-edge-robot architecture.

A training-free visual distillation method boosts VLA model performance in cluttered environments by over 34%, proving that targeted noise reduction is more effective than brute-force scaling.

Robots can now loosen screws with human-level dexterity thanks to a new framework that combines haptic estimation, online planning, and adaptive stiffness control using a parameterized Equilibrium Manifold.

A simple, low-cost smart waste bin design achieves touch-free operation using commodity STM32 microcontrollers and ultrasonic sensors.

By fusing language model reasoning with diffusion-based trajectory generation, KnowDiffuser leapfrogs existing autonomous driving planners on the nuPlan benchmark.

A new gripper design automates the tedious and injury-prone task of opening sterile medical pouches, freeing up nurses from a physically demanding, repetitive procedure.

A single meta-RL policy can now handle 66% mass variations and 70% rotor thrust losses in quadrotors, achieving zero-shot sim-to-real transfer for agile maneuvers.

Gaussian trajectory predictors often lie about their confidence, but a new loss function leveraging Kernel Density Estimation can make them more honest, leading to safer autonomous navigation.

By decoupling visual and motor information during pretraining, FutureVLA unlocks more effective visuomotor prediction for vision-language-action models, boosting performance without modifying downstream architectures.

Guaranteeing safety in diffusion-based trajectory planning is now possible by embedding a certifiable barrier function directly into the denoising loop, ensuring forward invariance and preserving the learned path geometry.

By jointly modeling video dynamics and actions, DiT4DiT achieves 10x sample efficiency and 7x faster convergence in robot policy learning, showing that video generation can be a powerful scaling proxy.

Forget fine-tuning: this method adapts robots to changing environments by learning a low-dimensional "Trend ID" embedding, keeping the core model fixed.

Robots can now scrape vials like a human chemist, thanks to a reinforcement learning policy that adapts force in real-time based on visual feedback.

Overcoming the data scarcity bottleneck in robotic arm-hand coordination, FAR-Dex achieves over 80% real-world success in fine-grained dexterous manipulation tasks.

Achieve efficient task execution in shared workspaces by interleaving scheduling and motion planning, using symbolic feedback to guide the scheduler towards motion-feasible solutions.

A nose-mounted microphone and vibration sensor combo unlocks robust, low-audibility speech interfaces for always-on AI interaction, even in noisy environments.

Bypass the need for extensive on-site data collection when deploying pre-trained robot models by visually prompting them to adapt to new scenes.

Autonomous vehicles can now better "see" the world even when cameras fail, thanks to a new method that fills in the blanks by leveraging spatial overlaps and learned semantic priors.

By converting point clouds into a format VLMs can understand, VLM-Loc significantly boosts text-to-point-cloud localization accuracy, outperforming prior methods that rely on shallower text-point cloud correspondences.

Fine-grained foot motion capture, a notoriously hard problem, gets a 30% accuracy boost by cleverly lifting 2D keypoints to 3D using motion capture data and contextual information, bypassing the need for direct image-3D annotation pairs.

Forget manual labeling: STONE offers a massive, automatically-labeled dataset for off-road robot navigation, unlocking scalable training for robust 3D traversability prediction.

Drones can now proactively navigate turbulent environments thanks to a fast wind-prediction framework that integrates geometric perception and local weather data.

Ditch the map: a diffusion model learns to plan UAV swarm trajectories directly from RGB images, enabling reactive and adaptive navigation in cluttered environments.

Human-in-the-loop learning can now boost dexterous manipulation VLA models by 25%, thanks to a new framework that smartly samples corrective actions and enables real-time intervention.

Humanoid robots can now walk robustly in the real world using only onboard sensors, thanks to a new diffusion policy that implicitly learns state estimation.

Forget hand-crafted heuristics: this new dynamics-aware policy learns to exploit contact forces in cluttered environments, outperforming traditional methods by 25% in simulation and showing impressive sim-to-real transfer.

Physics-based dynamics models can make or break sim-to-real reinforcement learning, boosting real-world success by 50% in industrial control tasks where simplified models fail.

Unlock real-time semantic SLAM and multimodal interaction with 3D Gaussian Splatting using X-GS, a unified and extensible open framework.

Forget hand-engineered reward functions: this method uses language models to learn factorized world states that generalize to new goals and environments, outperforming LLM-as-a-Judge in zero-shot reward prediction.

Offline RL can be made more robust to distribution shift by directly optimizing against worst-case transition dynamics within an uncertainty set, leading to policies that avoid unreliable out-of-distribution actions.

Ditch the latency tax of traditional scheduling: this new approach delivers data "just-in-time" for safety-critical systems, boosting performance without sacrificing reliability.

Imagine robots that can literally grow new sensors on demand, adapting to their environment in real-time through internal chemical reactions.

A caterpillar-inspired robot can now squeeze into tight spaces and "feel" its way around using artificial bristles, offering a cost-effective upgrade for existing robotic arms.

A 4B-parameter model outperforms Gemini-3-Pro in autonomous driving by incorporating physics-informed constraints and style-aware training, suggesting specialized models can surpass larger, general-purpose models in domain-specific tasks.

A complete, GPU-accelerated bimanual mobile manipulation platform can be built for under $1300, opening up robotics research and education to a wider audience.

Achieve robust locomotion for multi-legged robots on rough terrain with a surprisingly simple, decentralized control architecture that blends event-driven and CPG-based approaches.

By incorporating language guidance into federated learning, SurgFed tackles the long-standing problem of tissue and task heterogeneity in surgical video understanding, leading to improved segmentation and depth estimation across diverse surgical settings.

Finally, a GelSight-style sensor that doesn't force you to choose between pre-contact vision and high-fidelity tactile sensing.

Ditch the flat scene graphs: TopoOR models surgical environments as higher-order topological structures, unlocking superior performance in safety-critical tasks by preserving complex relationships and multimodal data.

Zero-shot robotic manipulation is now within reach: TiPToP matches a 350-hour fine-tuned model without *any* robot data.

Task demands in remote AR collaboration dictate how much network delay users can tolerate before perceived fluency breaks down, paving the way for adaptive systems.

Collect high-quality robot manipulation data anywhere with TRIP-Bag, a teleoperation system that fits in a suitcase and sets up in under 5 minutes.

Forget waiting hours: this MORL framework achieves 270x speedups on robotics tasks thanks to GPU-native parallelization.

Unlock the power of web videos for embodied AI: implicit geometry representations let agents learn to navigate from real-world room tours without relying on fragile 3D reconstruction.

Ignoring CSI phase information in robotic activity recognition is a mistake: fusing it with amplitude data in a novel gated BiLSTM architecture significantly boosts accuracy and robustness.

By representing visual inputs as 3D Gaussian primitives, GST-VLA unlocks a new level of geometric understanding for vision-language-action models, leading to substantial performance gains in robotic manipulation tasks.

Stop letting simulator errors in critical regions derail your policies: Sim2Act aligns surrogate fidelity with downstream decision impact, leading to more stable and robust decision-making.

Overcoming GPS trajectory matching limitations in dense urban areas with low-frequency data is now more achievable thanks to enhanced spatio-temporal matching strategies.

See how tweaking your 3D print orientation and parameters *before* printing can slash surface roughness, thanks to this interactive roughness prediction tool.

A robot can now achieve 90% success in peg-in-hole tasks, even with only 0.1mm clearance, by intelligently fusing vision and tactile feedback when visual occlusion occurs.

Combining pre-trained and custom neural networks with data augmentation and transfer learning yields a robust autonomous driving system capable of accurately perceiving and reacting to its environment.

Reconstructing and simulating wind-driven dynamics from video is now possible with a new differentiable framework that enforces fluid dynamics laws.

Robots can now recover from failures during manipulation tasks by explicitly tracking progress against spatial subgoals, without needing extra training data or models.