AI Worker FFW-SG2
About
About AI Worker FFW-SG2
AI Worker FFW-SG2 is ROBOTIS's mobile dual-arm platform for teaching and executing manipulation tasks. Its two seven-joint arms and adaptive parallel grippers sit on a lifting torso above a three-wheel swerve-drive base. A separate leader device records human demonstrations, while ROS2 tools connect data collection, learning, navigation and operation on the robot.
The SG2 combines a head stereo camera, two close-range gripper cameras, two lidar scanners and an IMU with NVIDIA Jetson AGX Orin computing. Each arm is rated for three kilograms, with a five-kilogram peak that depends on posture and motion. The standard grippers are included in the published configuration; advanced finger actuators remain under development. SG2 is the mobile model in the documented lineup. The similarly named BG2 has no mobile drive and uses an external power supply.
Capabilities
Source-backed capabilities
- Control modes: Leader Follower Teleoperation
- Remote and shared control
Claimed - Mirror an operator’s arm movements, actuate the grippers and drive the swerve base using the leader controls.
Source
Observed use
Exact-product configurations are shown without inferring supply, purchase, customer status, standard fit, deployment or partnership.
- ROBOTIS RH-P12-RN
Reference Design - Mounted - Physical - ROBOTIS lists RH-P12-RN two-finger grippers as standard equipment on the FFW-SG2 AI Worker.
Source
Research
- ROBOTIS AI Worker ROS 2 platform and simulation packages
Research Publication - Official ROS 2 packages for controlling AI Worker robots, with robot descriptions, sensor interfaces, navigation and teleoperation components.
Source - ROBOTIS Cyclo Intelligence
Research Publication - Cyclo Intelligence provides demonstration recording, dataset conversion, policy training, inference and robot execution infrastructure for ROBOTIS AI Worker. The source includes FFW-SG2 configuration, a robot description and an example behavior tree for the orchestration stack.
Source - ROBOTIS Cyclo Control
Research Publication - Cyclo Control provides constrained motion control, kinematics, QP optimization and retargeting packages for ROBOTIS robots. Its FFW-SG2 descriptions and semantic models support MoveL, MoveJ, bimanual control and ROS 2 controller integration.
Source - ROBOTIS Cyclo Lab simulation environments
Research Publication - Cyclo Lab provides Isaac Lab environments, robot assets and reinforcement/imitation learning configurations for ROBOTIS robots. Exact configurations include FFW-SG2 pick-and-place and AI Sapiens K1 locomotion and motion imitation.
Source
Context
Comparable Systems
- Moz1
Spirit AI - Semi Humanoid
- YMBot D
Yunmu Intelligent Manufacturing - Semi Humanoid
- YMBot D2
Yunmu Intelligent Manufacturing - Semi Humanoid
- AGIBOT G1 MAX
AgiBot - Semi Humanoid
- HMND 01 Alpha Wheeled
Humanoid - Semi Humanoid
- KUAVO 5W
Leju Robotics - Semi Humanoid
