Home Bots & Business Pudu Gives Service Robot Arms

Pudu Gives Service Robot Arms

by Marco van der Hoeven

Pudu Robotics has introduced the FlashBot Arm, a semi-humanoid service robot designed for commercial use. Developed by the company’s research division, Pudu X-Lab, the FlashBot Arm builds on the features of its predecessor, FlashBot, by integrating robotic manipulation capabilities with autonomous delivery and interaction functions supported by embodied artificial intelligence.

According to Pudu Robotics, the FlashBot Arm is equipped with two robotic arms, each with seven degrees of freedom, and hands with 11 degrees of freedom, which the company refers to as PUDU DH11. This design enables the robot to perform a range of tasks such as pressing elevator buttons, picking up and transporting items, and interacting with physical environments in a way that does not require changes to existing infrastructure. The robot’s operational reach extends up to two meters.

The platform also incorporates simultaneous localization and mapping (SLAM) technologies—specifically visual SLAM and laser SLAM—to generate 3D maps, navigate autonomously, and avoid obstacles. Its closed delivery compartment is intended to secure items during transport across environments such as hotels, office buildings, restaurants, and healthcare facilities.

In addition to its physical capabilities, the FlashBot Arm includes a large language model-based system designed to enable conversational interactions and task planning. This feature allows the robot to understand verbal commands, decompose tasks, and coordinate with other robotic units within the same environment.

Pudu Robotics reports that the robot employs whole-body control, which synchronizes movement across its base, arms, and hands, enabling real-time task adjustments. Safety mechanisms include RGBD cameras, LiDAR, panoramic vision, pressure-sensitive skin, and dynamic motion planning algorithms that allow the robot to halt upon detecting an obstacle. User interaction is facilitated through multimodal input methods including voice, gesture, and facial recognition. The company states that these features aim to make the robot more accessible in everyday commercial settings.

 

Misschien vind je deze berichten ook interessant

preload imagepreload image