Motor- and kinematics-based force estimation gives surgical robots real-time visual feedback for safer, more precise control in body cavities.
Semantic navigation, pose estimation, and coordinated arm-base motion improve autonomous grasping flexibility and success in dynamic environments.
A modular power pack and unified connector deliver electrical power and fluid reliably to wearable exoskeleton actuators while keeping the system portable.
An actuated tube end effector holds and obstructs an inverted fruit tube to retain multiple items with lower impact and faster packing.
3D vision, ArUco tags, and visual servo automate micromanipulator tip initialization, cutting manual setup time and tip damage risk.
Dynamic resampling of 3D Gaussian particles helps robots track and predict deformable object states under weak visual cues and complex deformations.
Mounting the robot and working machine on one movable stand helps keep workpieces aligned for smoother gripping, transfer, and placement.
Real-time force feedback lets a robotic sanding head follow toolpaths accurately while improving surface precision and throughput.
Face mapping, formula dosing, and airbrush spraying automate makeup application to improve hygiene, color matching, and application speed.
Natural language coordination lets AI assign capability-matched robots faster, improving flexible multi-robot task execution and debugging.
A lever arm, pivoting fulcrum, and rotating cam grip flexible palm stalks securely, cutting the force needed for complete seed pod extraction.
Conversation history, emotion, and health analysis let a care robot prioritize personalized services and respond more naturally.
Timed prompt generation helps a care robot match user needs from sensor data while limiting costly language model calls.
Uses LLM intent extraction and 3D semantic scene graphs to help robots complete tasks reliably in unstructured environments.
An on-board nanofactory and symbolic AI governance let humanoid robots self-repair, recycle parts, and maintain sovereign compliance in extreme environments.
Pivoting clamps and a support rope center limbs of different diameters to maintain exoskeleton joint alignment and reduce parasitic forces.
Gaussian Splatting and forward occupancy bounds make radiance-field robot planning less conservative while preserving real-time collision safety.
Optical multi-point correction maps image coordinates to pipette position, reducing mechanical deviation and improving biological particle sorting accuracy.
Real-time leg-speed estimation and center-of-mass trajectory control keep a bipedal exoskeleton stable during acceleration and deceleration.
Human-specific feature extraction and deductive sensor fusion help distinguish people from AGVs, reducing unnecessary safety stops.
Lower-body measurements tailor joint angles and stride to reduce discomfort, prevent foot dragging, and improve gait training.
A monolithic pleated elastomer actuator enables contraction and extension while simplifying manufacture and reducing stress concentrations.
A passive spring-linkage counterbalance distributes one constant force across translational and rotary joints to cut weight and adapt to gravity changes.
Palm sensing and a trained model let a humanoid robot reproduce skilled cloth gripping and sewing motions with higher work accuracy.
Runtime screenshots and user-marked UI errors feed CV model retraining, improving RPA component recognition across interface changes.
A nested motor, gear, and decelerator axis layout shrinks the robot arm while improving assembly and reducing noise.
Auto-labeled vision classifiers let modular robots detect bowls independently, improving food assembly accuracy and throughput under varying conditions.
Position-adjustable bevel gears let robot joint meshing be tuned outside the arm, improving assembly access, transmission efficiency, and compactness.
A learned reference-point approach helps robots interpret ambiguous instructions, identify target objects, and generate grasp commands in cluttered scenes.
Hierarchical passivity constraints modulate energy flow to keep robot-human interaction stable and comfortable despite delays and human variability.
Pre-grasp perception and 3D feature references improve pose tracking of robot-carried objects after shifting, reducing handling errors and collisions.
A barbell model decouples front and rear leg height control, improving quadruped stability and response under changing body pitch.
Valve-regulated air compression stores and smooths recoil assistance in a piston, enabling adjustable exoskeleton force without tanks or jerky springs.
A friction gradient on collapsible gripper protrusions enables stable grasping, controlled slippage, and object property sensing.
A two-axis rotational end-of-arm tool changes pose to handle different objects without tool swaps, cutting delay, complexity, and storage needs.
Depth and color image predictions rank candidate arm actions to separate tightly stacked objects and improve grasp success with fewer collisions.
Autonomous tank seal inspection uses onboard sensors, imaging, and AI to detect seal defects faster while reducing human exposure to VOC hazards.
Two opposing robot arms reposition and lift objects through pushing and environmental support, avoiding complex grippers and delicate-item damage.
A constrained third limb and unconstrained limb group preserve stiffness while expanding workspace for efficient machining of large parts.
A semantic engine analyzes multi-device data to pre-route robotic actions toward affirmative outcomes with less control complexity.
Multiple robotic end effectors identify and open ophthalmic surgical packages automatically, reducing manual setup effort and item mix-ups.
A smaller on-device LLM lets companion robots answer locally, reducing network load, speeding response, and keeping user data on the terminal.
Magnetic coupling moves pharmaceutical containers through a sealed carrier body, reducing abrasion and contamination while keeping high processing rates.
Independent rotating and extending tool arms let one robotic end effector fasten and weld in parallel, cutting task time and boosting throughput.
A conveyor and swiper arm coordinate object release and placement to prevent drops and improve packing density in containers.
Direct constraint feedback to both robot and operator cuts teleoperation latency, improving collision avoidance and control precision.
A spring-loaded carriage and belt drive vary joint stiffness through lever-arm position, keeping robotic links compact while decoupling adjustment from load.
A movable coupler and lock let the gripper self-adjust position and tilt, improving insertion accuracy while reducing sensing and control load.
Bellows-type hydraulic actuators give robotic fingers compliant, sensor-guided control for delicate in-hand manipulation without damaging objects.
Remote hydraulic actuation cuts manipulator mass to preserve haptic feedback while enabling multi-DOF needle control in MRI settings.