New
My Cart
0
item(s)
-
€0.00
Get a quote in only 3 steps!
1. Add the items you are interested in to your cart
2.Go to your cart
3.Click on GENERATE A QUOTE
Tous vos devis sont accessibles à partir de votre compte client.
Robotis AI Worker Semi-Humanoid Robot
Robotis | A-000000-08094
Price upon request
AI WORKER is a “Physical AI” robotic platform combining dual 7-DOF arms, RGBD perception, and onboard Jetson AGX Orin computing, designed to learn tasks through demonstration and execute them in real-world conditions (assembly, inspection, logistics).
Payment methods
(see T&C)
Debit card
PayPal
Payment by invoice
Wire transfer
ROBOTIS AI WORKER: semi-humanoid robot for manipulation and AI in industrial environments
- Semi-humanoid robot focused on Physical AI, designed for industrial environments.
- Learning by demonstration: data collection via teleoperation, then improvement through imitation learning and reinforcement learning to execute robust tasks.
- Use cases: assembly, inspection, logistics (dynamic environments, fine and repeatable tasks).
- Complete architecture: base (depending on version) + dual 7-DOF arms + grippers + head + lift, for manipulation and interaction.
- Sensing for perception: multi RGBD cameras (and LiDAR + IMU on the mobile version) for navigation, obstacle avoidance and manipulation.
- Onboard computing: NVIDIA Jetson AGX Orin 32GB to run AI workloads and complex processing.
- Open-source platform: published software resources and models; AI Worker / AI Manipulator under the Apache-2.0 license.
Two versions available: mobile (FFW-SG2) or stationary (FFW-BG2)
AI WORKER comes in two “Follower” versions depending on your use case.
FFW-SG2 is the mobile version: omnidirectional Swerve Drive base, additional sensors (LiDAR, IMU) and battery power to operate in changing environments.
FFW-BG2 is a stationary version (no mobile base), designed for a fixed workstation use with AC power and direct Ethernet integration. Both share the same manipulation core: dual 7-DOF arms, grippers, head, lift and an Jetson AGX Orin 32GB computer.
| Feature | FFW-SG2 (Mobile) | FFW-BG2 (Stationary) |
| Dimensions (L × W × H) | 604 × 602 × 1,623 mm | 604 × 564 × 1,607 mm |
| Weight | 90 kg | 85 kg |
| Degrees of freedom (DOF) | Total 25 DOF Arms 7 DOF × 2 · Gripper 1 DOF × 2 · Head 2 DOF · Lift 1 DOF · Mobile 6 DOF | Total 19 DOF Arms 7 DOF × 2 · Gripper 1 DOF × 2 · Head 2 DOF · Lift 1 DOF |
| Arm reach | 641 mm (to wrist) + hand | 641 mm (to wrist) + hand |
| Arm payload (nominal) | 3.0 kg (single arm) / 6.0 kg (dual arm) | 3.0 kg (single arm) / 6.0 kg (dual arm) |
| Arm payload (peak) | 5.0 kg (single arm) / 10.0 kg (dual arm) | 5.0 kg (single arm) / 10.0 kg (dual arm) |
| Base / mobility | Swerve Drive | None (fixed workstation) |
| Travel speed | 1.5 m/s | — |
| Power | Battery 25V, 80Ah (2,040Wh) | AC powered via SMPS: 24VDC, 80A, 1,920W |
| Sensors | RGBD × 3 · LiDAR × 2 · IMU | RGBD × 3 |
| Host interface | Ethernet (Wi-Fi router up to 1 Gbps) | Ethernet (direct connection to host PC) |
| Internal communication | RS-485 · 4 Mbps | RS-485 · 4 Mbps |
| Computer | NVIDIA Jetson AGX Orin 32GB | NVIDIA Jetson AGX Orin 32GB |
| Software | ROS 2 · Python · C++ · Web UI | ROS 2 · Python · C++ · Web UI |
| Operating temperature | 0 to 40°C | 0 to 40°C |
| Exterior materials | Aluminum, plastic | Aluminum, plastic |
To select the right version, start by qualifying your environment: do you need autonomous movement (navigation, obstacle avoidance, multi-station tasks)? If yes, FFW-SG2 (Swerve Drive, LiDAR, battery) is the best fit. If your application is a repeatable task on a fixed workstation (inspection cell, assembly station), FFW-BG2 is often easier to integrate (AC power, direct Ethernet).
Applications and project types (Physical AI, imitation learning, manipulation, mobility)
AI WORKER targets teams (industry, R&D, laboratories) that want to develop and validate “real-world” robotic skills based on learning by demonstration: the robot reproduces a learned gesture, then becomes more robust thanks to imitation learning and reinforcement learning.
This approach is especially relevant for tasks such as assembly, inspection and certain logistics scenarios in dynamic environments.
Sensors and perception (RGBD, LiDAR, IMU): what it changes for manipulation
AI WORKER relies on a multi-camera system to combine global understanding (navigation, localization, avoidance) and short-range perception (object grasping, manipulation refinements).
The head camera is a Stereolabs ZED Mini (stereo RGBD + 6DoF IMU) for wide-area perception, and the hand cameras are Intel RealSense D405 (x2) integrated at the grippers for close-range depth, useful for grasping and precision manipulation.
The FFW-SG2 version adds 2 LiDARs and an IMU to strengthen perception while moving.
Grasping: standard RH-P12-RN gripper
The standard RH-P12-RN gripper is a 2-finger / 1-DOF robotic hand with an “adaptive gripping” design (passive joints) to grasp objects of various shapes. It uses a 10W DC motor and supports torque control and current-based position control. The fingertips are interchangeable.
Software ecosystem (ROS 2, Physical AI tools, simulation, models & datasets)
ROBOTIS highlights a toolchain to move quickly from prototype to reproducible experiments: ROS 2 support, Python/C++ development, a Web UI, “Physical AI” tools, and simulation environments (models for Isaac Sim and MuJoCo). Additional resources are available to speed up onboarding (ROS 2 packages, models, datasets, and Docker images).
Technical specifications – ROBOTIS AI WORKER
| Versions | FFW-SG2 (mobile) / FFW-BG2 (stationary) |
| Arms | 7 DOF × 2 |
| Grippers | 1 DOF × 2 (RH-P12-RN, standard) |
| Head | 2 DOF |
| Lift | 1 DOF (stroke 0 to 500 mm) |
| Arm reach | 641 mm (wrist) + hand |
| Payload (nominal) | 3.0 kg (single arm) / 6.0 kg (dual arm) |
| Payload (peak) | 5.0 kg (single arm) / 10.0 kg (dual arm) |
| Sensors | RGBD × 3 (and LiDAR × 2 + IMU on FFW-SG2) |
| Computer | NVIDIA Jetson AGX Orin 32GB |
| Interfaces | Ethernet (Wi-Fi via router on FFW-SG2) |
| Internal communication | RS-485 · 4 Mbps |
| Software | ROS 2 · Python · C++ · Web UI |
| Temperature | 0 to 40°C |
| Exterior materials | Aluminum, plastic |
Cameras – perception (head & hands)
| Head camera | Stereolabs ZED Mini (stereo RGBD + 6DoF IMU) |
| FOV (head) | 102°(H) × 57°(V) |
| Max resolution (head) | 2208 × 1242 |
| Depth range (head) | 0.1 m to 9 m |
| Hand cameras | Intel RealSense D405 × 2 (stereo RGBD) |
| FOV (hands) | 87°(H) × 58°(V) |
| Max resolution (hands) | 1280 × 720 |
| Depth range (hands) | 7 cm to 50 cm |
Gripper – RH-P12-RN (standard)
| Type | 2-finger robotic hand, 1 DOF, “adaptive gripping” |
| Motor | DC 10W |
| Control | Torque control + current-based position control |
| Weight | 500 g |
| Gripper payload | 5 kg |
| Fingertips | Detachable and customizable tips |
Resources – ROBOTIS AI WORKER
FAQ – ROBOTIS AI WORKER
What is the key strength of AI WORKER?
A “Physical AI” platform ready for real-world deployment: dual arms + multi-sensor perception + onboard computing, with a demonstration-learning workflow (teleoperation → data → model → execution).
Which version should I choose: FFW-SG2 or FFW-BG2?
FFW-SG2 if you need mobility (navigation, multi-station work, variable environment). FFW-BG2 if your application is a fixed workstation (inspection/assembly station, direct Ethernet integration).
Is the RH-P12-RN gripper suitable for industrial manipulation?
Yes for versatile utility grasps: two fingers, adaptive design, interchangeable tips.
What software stack is provided?
ROS 2 support, Python/C++ development, and a Web UI; simulation resources and related packages.
AI WORKER is a “Physical AI” robotic platform combining dual 7-DOF arms, RGBD perception, and onboard Jetson AGX Orin computing, designed to learn tasks through demonstration and execute them in real-world conditions (assembly, inspection, logistics).
Customers who viewed this item also viewed
6,427.20 €
VAT incl.
9,971.52 €
VAT incl.