Solutions & Services Overview
In the surging wave of next-generation robotics and smart manufacturing, enterprises face two critical challenges: efficiently training complex AI models and deploying server-grade performance to the edge for real-time decision-making. NexCOBOT delivers an integrated AI Training Server and AI Edge Deployment Solutions, establishing a highly efficient, robust, and coherent closed-loop ecosystem for your Physical AI journey.
AI Edge Controller : NexCOBOT has partnered with NVIDIA to launch a development kit and edge computing system featuring the cutting-edge NVIDIA Jetson Thor accelerator. This solution brings server-class AI performance directly to mission-critical edge applications, including humanoid and quadruped robotics, mobile manipulation, and smart manufacturing. The technology facilitates real-time multi-modal data analysis and ensuring superior decision-making.
AI Training Server : Designed to accelerate robotic simulation, motion planning, and digital twin development through full integration with NVIDIA Omniverse, CuMotion, FoundationPose, and ROS (Robot Operating System). It provides a unified environment where developers can design, simulate, and optimize intelligent robots before real-world deployment.
- Environment Setup & System Integration
- Install latest Isaac Sim
- Install and configure Python3, Jupyter Notebook, and VS Code
- Deploy ROS2 Humble and MoveIt2
Install ISAAC ROS modules:
- isaac_ros_rtdetr
- isaac_ros_foundationpose
- isaac_ros_cumotion
- Deploy NVIDIA TAO Toolkit environment
- Training & Documentation
- Isaac Manipulator training guide
- RT-DETR training tutorial
- Sample Code Delivery
- Python script for synthetic image dataset generation
- Python script for COCO dataset conversion
- Jupyter Notebook for RT-DETR training
|
