新闻
Chinese medical-device company Aixam announced that its M1 autonomous blood-collection robot received formal medical-device registration and is entering the market. The system combines vein sensing and feedback, autonomous decision-making, and motion control to locate veins in 3D. Aixam reports that M1 achieved 95% success in rat-tail vein sampling experiments, compared with a cited prior result of 83.9%.
Robot Newspaper论文
OpenTac combines a scalable resistive tactile skin with a CAD tool that generates fabrication files for custom 2D and developable 3D surfaces. Its structured overlay encodes normal force and local shear through taxel responses. The team designed and fabricated more than 30 sensors and demonstrated tactile feedback control and manipulation-policy learning on four real-world robot embodiments. Adaptive readout reached kilohertz rates and was 19.2× faster than full-array scanning at 16 of 1,024 taxels per update.
arXiv论文
Progress Probes trains linear probes on activations from the open-source π0.5 vision-language-action model, using successful demonstrations to estimate task progress. At runtime, the monitor flags executions whose estimated progress lags or deviates from the expected trajectory. Across simulated and real-world manipulation tasks, the method was competitive with or outperformed VLM-based monitors and methods trained with failure data, without requiring an external foundation model.
arXiv cs.RO论文
VecTile is a 6 mm-thin, flexible tactile array that uses a piezoresistive layer and silicone nubs to distinguish normal from shear forces. A 4×4 array costs less than $35; in single-taxel tests, normal- and shear-force errors were 0.39 N and 0.16 N RMS. The authors demonstrate balloon-safe scalpel contact and improved Diffusion Policy-guided power-plug insertion.
arXiv cs.RO论文
FocusVLA is a stage-aware multi-view routing method for robotic manipulation. It predicts future interaction intent from action history and uses that prediction to dynamically coordinate global and wrist/local camera views during action generation. On RoboTwin 2.0, it improves average success by 10.10 percentage points over the strongest baseline, π₀.₅; in real-world evaluation, reported success rises from 71.7% to 85.0%.
arXiv cs.RO论文
GelSlim 5.0 uses one fixed fingertip camera and a UV-responsive contact layer to capture visual and tactile images sequentially by switching illumination. The VisTacFusion transformer estimates planar object pose and contact depth and surface normals, with modality dropout supporting paired or single-modality inputs. On real validation data from 12 known objects, paired inputs reduced depth MSE by 61.2% versus tactile-only inference and 93.7% versus visual-only inference.
arXiv cs.RO演示
Introduces HX Friends robotic hand and gripper designs with different finger counts, built on DYNAMIXEL. The lineup includes a single-actuator gripper and a single-finger design.
ROBOTIS新闻
ETH Zurich researchers turned an off-the-shelf robotic hand into a self-contained mobile manipulator, adding an 80-gram backpack with a battery, IMU, and Raspberry Pi Zero. Reinforcement learning trained it to crawl and perform separate skills including righting itself, pressing keys, and pushing objects. In tests, it crossed 14 surfaces at an average 9 cm/s and pushed a 41-gram cube to targets up to 40 cm away in 15 consecutive trials.
IEEE Spectrum演示
Agility's Digit rearranges a room with varied objects, demonstrating manipulation for mixed loads and changing layouts.
Agility演示
Agility's Digit robot rearranges assorted objects in a room, demonstrating generalized manipulation.
Agility演示
Unitree presents UnifoLM-OminiA-0.3 for whole-body mobile manipulation and home-care and wellness tasks.
Unitree Robotics演示
Agility's Digit is shown on grill duty; the description highlights precise force control.
Agility