- Humanoid's HMND 01 Alpha wheeled humanoid was trialed at Siemens' Erlangen electronics factory for autonomous tote-handling logistics.
- Trial metrics: 60 tote moves/hour, uptime >8 hours, and pick-and-place success rates above 90%.
- Humanoid integrated NVIDIA's physical AI stack including Jetson Thor, Isaac Sim and Isaac Lab for edge compute, simulation and RL training.
- Simulation-first hardware and training reduced prototype development from 18-24 months to 7 months.
Overview
Siemens and Humanoid trialed the HMND 01 Alpha humanoid robot in Siemens' Erlangen electronics factory to perform autonomous tote-handling logistics using NVIDIA's physical AI stack.
Deployment results
The robot achieved 60 tote moves per hour, operated autonomously for more than eight hours of uptime, and maintained pick-and-place success rates above 90% while interacting with human operators.
System integration
Integration with Siemens Xcelerator provided digital-twin capabilities, PLC-robot interfaces, fleet management, industrial communications and drives to enable real-time data exchange and synchronized workflows with other equipment.
AI and development
Humanoid used NVIDIA Jetson Thor for edge inference, Isaac Sim for simulation and Isaac Lab for reinforcement-learning policy training; a simulation-first approach reduced prototype development time from a typical 18–24 months to seven months.
Platform design
The HMND 01 combines omnidirectional wheeled mobility with advanced manipulators and a proprietary AI framework designed to operate in human-centric factory environments and adapt to diverse tasks.
