How far is demo capability from deployment economics?
THEME RESEARCH
Embodied AI & Robotics
Will value accrue first to robot data, edge compute, components, or operators that can generate repeat orders?
Paid deployments, intervention rate, uptime, and cost per completed task
01TOP SIGNAL
Full text reviewed
Read evidence →Humanoid hands learned three real world tasks from modest data
The takeaway: The recipe transfers three vision-based dexterity tasks to a real humanoid, but reported success is 52.5-80% and still far from general-purpose manipulation.
02
Long VLA kept robots on track through multi step jobs
Long-VLA improves multi-step manipulation by separating movement from interaction inside one model, but dataset labeling is manual and longer sequences still fail.
Full text reviewed
Tesla
Proceedings of Machine Learning Research
03A robot learned to open doors without damaging either side
SafeDiff uses real-time force signals to correct a vision-planned path and transfers with limited real data, but it has only been tested on one constrained task with a gripper.
Full text reviewed
Tesla
Proceedings of Machine Learning Research
04Hand over hand teaching cut dexterous robot training time in half
KineDex lets a person physically guide a robot hand, collecting tactile demonstrations faster and improving contact-heavy task success, but the setup is not yet scalable to two hands.
Full text reviewed
Tesla
Proceedings of Machine Learning Research
05A 3D world model learned how rigid and soft objects move together
ParticleFormer predicts how rigid, soft, and flexible objects move well enough to improve model-based control, but each scene needs its own model and segmentation mistakes flow through the system.
Full text reviewed
Tesla
Proceedings of Machine Learning Research
01TOP SIGNAL
Source metadata only
Read evidence →Agtonomy releases new autonomous multi-point turning features
The takeaway: Agtonomy has added autonomous multi-point turning and enhanced field data collection to its commercial off-road platform. The post Agtonomy releases new autonomous multi-point turning features appeared first on The Robot Report .
02
A robot model still needs body-specific work before it can move to a new machine
Robot foundation models (RFMs), including vision-language-action (VLA) policies, are often discussed through a scaling view: more data, larger models, and broader benchmarks should improve generalization. In robotics, however, a model can generalize while work still remains before it can run on a robot with a particular body. The work required differs across methods and target robots, and those differences affect practical deployment. We call the gap between reusable models, representations, or data and their use in execution on the target robot the embodiment gap. This survey examines what can be reused across robot embodiments and what must still be implemented on a new robot. We place existing methods on a two-axis map that shows the type of shared structure and the stage at which adaptation is needed for execution on the target robot. We then examine recent work through three overlapping research directions: sharing semantics and perception, sharing robot data and interfaces, and learning correspondence across embodiments. We also propose a reporting framework for adaptation work that success rate alone does not reveal. The framework identifies the work that should be checked when comparing cross-embodiment learning and highlights work that remains on a new robot and questions for future study.
Abstract screened
arxiv-cs.RO
03Kollmorgen to give a joint-by-joint guide to humanoid motion at RoboBusiness
The talk at RoboBusiness will outline the unique demands of each body zone and the challenges of scaling for production. The post Kollmorgen to give a joint-by-joint guide to humanoid motion at RoboBusiness appeared first on The Robot Report .
Source metadata only
rss:www.therobotreport.com
04Pudu Robotics launches new MP2000 autonomous forklift
Pudu Robotics has launched the MP2000, an AI-native autonomous robot designed to transport 2,000 kg loads across complex warehouse environments. The post Pudu Robotics launches new MP2000 autonomous forklift appeared first on The Robot Report .
Source metadata only
rss:www.therobotreport.com
05Serve Robotics to deploy its autonomous delivery robots with Grubhub
Serve's robot delivery will be available from more than 100 participating Grubhub merchants in Chicago and nearly 200 in Los Angeles. The post Serve Robotics to deploy its autonomous delivery robots with Grubhub appeared first on The Robot Report .
Source metadata only
rss:www.therobotreport.com
01TOP SIGNAL
Abstract screened
Read evidence →Robots completed the task and still damaged the object
The takeaway: Physical interaction quality is central to deformable-object manipulation, yet most benchmarks evaluate task success alone. A policy may complete the task while allowing slip or causing excessive compression. A primary bottleneck is the absence of visuo-tactile datasets that pair policy-visible contact observations with independent physical ground truth over complete tasks. We introduce SoftVTBench, a visuo-tactile dataset for physical-interaction-aware deformable-object manipulation. It contains 4,000 expert demonstrations and more than 50 assets, including volumetric deformable objects and visually matched rigid twins. At 20 Hz, each episode synchronizes multi-view RGB, dual-finger tactile RGB and marker motion, proprioception, language, and binary and continuous gripper actions, alongside evaluator-only finite-element (FEM) states. Building upon this dataset, we establish a closed-loop benchmark that uses fixed object-specific calibration to define the Deformation-aware Success Rate (DSR), which counts a rollout as successful only when it completes the task and keeps peak normalized deformation within tolerance. Across Diffusion Policy, $π_{0.5}$, and FastWAM, all 12 in-distribution configurations contain successful rollouts that violate the deformation tolerance, accounting for 0.7--24% of each configuration's successes. Under distribution shift, visuo-tactile variants achieve higher task success in all six policy--suite comparisons and higher DSR in five, whereas their in-distribution benefits are mixed. These results show that making touch available does not by itself ensure effective multimodal fusion. SoftVTBench therefore provides a common visuo-tactile resource for studying not only whether a policy succeeds, but how it physically interacts with deformable objects and when touch improves that interaction.
02
Multimodal Rapport Estimation in Real-World HRI
Evaluating interaction quality in real-world HRI is an important challenge. If interaction quality can be estimated reliably, the results can be used to improve dialogue strategies and ultimately enable robots to adapt their behavior autonomously. However, existing automatic evaluation methods have been developed primarily in controlled laboratory settings, and it remains unclear whether they can be directly applied to real-world environments, where users are free to disengage and multi-party participation may arise naturally. In this study, we investigate the automatic estimation of third-party-rated rapport scores using 62 sessions of multimodal recordings collected in a Japanese drugstore. We compare zero-shot LLMs, pretrained text, audio, and visual models, and their prediction-level fusion. The results show that, in real-world HRI, zero-shot LLMs achieve strong performance, while audio and visual models tend to provide complementary information. In particular, Gemini 2.5 Flash performs strongly as a single model, and a fusion model combining Gemini (text) with HuBERT and V-JEPA performs best overall. Further analyses showed that estimation performance varied across interaction-duration and group-size conditions. These findings suggest that rapport estimation in real-world HRI requires evaluation and model design that account for contextual variability beyond that assumed in laboratory settings.
Source metadata only
arxiv-cs.RO
03HarvestPoint-ACT: Explicit Target Selection and Harvest-Point Conditioning for Robotic Fruit Harvesting under Occlusion
End-to-end imitation learning avoids hand-made robot motion for approaching and grasping, but the policy must still decide which fruit to pick and where to close the gripper. Occlusion can make the policy lose the selected fruit during harvesting, and the correct closing point is difficult to infer from pixels alone. This paper presents HarvestPoint-ACT, which makes both decisions explicit in perception and provides them to the policy. An instance segmentation front end with a keypoint branch predicts a mask and a harvest point for each visible fruit, where the harvest point specifies the location to close the gripper. A scheduler ranks detected candidates by occlusion and travel distance and selects one target. After each attempt, it redetects and reranks the candidates because the canopy may have changed. The selected fruit is encoded for an action chunking transformer as an eight-dimensional state, containing the absolute harvest point, the vector from the gripper to that point, a validity flag, and a confidence score. When the selected fruit is temporarily undetected, the system retains the last harvest point estimate in the robot base frame and marks it as stale, and aborts the attempt if the loss persists. On a canopy mock-up, HarvestPoint-ACT achieves a success rate of 88%, and of 75% under heavy occlusion.
Source metadata only
arxiv-cs.RO
04Backward Layout Search for Sequence-Constrained Robotic Assembly
Robotic assembly layout planning must determine the assembly site and the initial pose of each part while ensuring collision-free execution of a prescribed assembly sequence. This problem is challenging because the obstacle environment changes after each assembly step, and unassembled parts re maining in the workspace may block robot motions. We observe that the feasibility of each assembly step depends only on the initial poses of the current and later-assembled parts. Based on this dependency, we propose Backward Layout Search (BLS), which assigns initial part poses in reverse assembly order. Each expansion performs geometric, kinematic, grasp, and prescribed-motion checks, while collision masks and candidate set filtering remove infeasible initial part pose candidates. Promising partial layouts are retained through beam selection, and complete layouts are validated by full motion planning in forward assembly order. Experiments on five assembly models show that BLS produces collision-free executable layouts and reduces step evaluations and search time compared with a matched forward search.
Source metadata only
arxiv-cs.RO
05DevGRU: Depth-guided Visual Navigation using a Collision-aware Recurrent Model
Existing visual navigation models often aim to develop foundation models that can generalize robot navigation across diverse platforms. However, many of these models are prone to collisions when deployed in complex indoor environments, particularly in structured layouts and narrow passages. To address this problem, we propose a depth image- and point-goal-conditioned navigation system, DevGRU. The proposed system employs an action predictor (AP) that generates collision-aware future trajectories, enabling effective avoidance of immediate obstacles. In conjunction with a collision predictor, the AP further compensates for errors accumulated in the goal pose estimation and proactively mitigates future deviations. To evaluate our method, we conducted experiments across nine different scenes and three state-of-the-art approaches - ViNT, NoMaD, and NavDP - as well as four additional variants of ViNT and NoMaD. In terms of navigation performance, DevGRU significantly outperforms ViNT and NoMaD by a large margin. In addition, the proposed model has a relatively small number of trainable parameters, resulting in the fastest inference time among the baselines, particularly outperforming NavDP by 7x in model size and 17x in inference time.
Source metadata only
arxiv-cs.RO