Intel Embodied Intelligence SDK

Embodied Intelligence SDK is a suite of intuitive, easy-to-use software stack designed to streamline the development process of Embodied Intelligence product and applications on Intel platform. The SDK provides developers with a comprehensive environment for developing, testing, and optimizing Embodied Intelligence software and algorithms efficiently. It also provides necessary software framework, libraries, tools, Best known configuration(BKC), tutorials and example codes to facilitate AI solution development.

Embodied Intelligence SDK includes below features:

• Comprehensive software platform from BSP, acceleration libraries, SDK to reference demos, with documentation and developer tutorials;

• Real-time BKC, Linux RT kernel and optimized EtherCAT;

• Traditional vision and motion planning acceleration on CPU, Reinforcement/Imitation Learning-based manipulation, AI-based vision & LLM/VLM acceleration on iGPU & NPU;

• Typical workflows and examples including ACT/DP-based manipulation, LLM task planning, Pick & Place, ORB-SLAM3, etc.

Software Architecture

Below picture is high level software architecture of Embodied Intelligence SDK:

This software architecture is designed to power Embodied Intelligence systems by integrating computer vision, AI-driven manipulation, locomotion, SLAM, and large models into a unified framework. Built on ROS2 middleware, it takes advantage of Intel’s CPU, iGPU, dGPU, and NPU to optimize performance for robotics and AI applications. The stack includes high-performance AI frameworks, real-time libraries, and system-level optimizations, making it a comprehensive solution for Embodied Intelligence products.

At the highest level, the architecture is structured around key reference pipelines and demos that demonstrate its core capabilities. These include Vision Servo, which enhances robotic perception using AI-powered vision modules, and ACT-based Manipulation, which applies reinforcement learning and imitation learning to improve robotic grasping and movement. Optimized Locomotion leverages traditional control algorithms like MPC (Model Predictive Control) and LQR (Linear Quadratic Regulator), alongside reinforcement learning models for adaptive motion. Additionally, the ORB-SLAM3 pipeline focuses on real-time simultaneous localization and mapping, while LLM Task Planning integrates large language models for intelligent task execution.

Beneath these pipelines, the software stack includes specialized AI and robotics modules. The vision module supports CNN-based models, OpenCV, and PCL operators for optimized perception, enabling robots to interpret their surroundings efficiently. The manipulation module combines traditional motion planning with AI-driven control, allowing robots to execute complex movements. For locomotion, the system blends classic control techniques with reinforcement learning models, ensuring smooth and adaptive movement. Meanwhile, SLAM components such as GPU ORB extraction and ADBSCAN optimization enhance mapping accuracy, and BEV (Bird’s Eye View) models contribute to improved spatial awareness. The large model module supports LLMs, Vision-Language Models (VLM), and Vision-Language-Action Models (VLA), enabling advanced reasoning and decision-making capabilities.

At the core of the system is ROS2 middleware and acceleration frameworks, which provide a standardized framework for robotics development. The architecture is further enhanced by Intel’s AI acceleration libraries, including Intel® OpenVINO™ for deep learning inference, Intel® LLM Library for PyTorch* (IPEX-LLM) for optimized large model execution, and compatibility with TensorFlow*, PyTorch*, and ONNX*. The Intel® oneAPI compiler and libraries offer high-performance computing capabilities, leveraging oneMKL for mathematical operations, oneDNN for deep learning, and oneTBB for parallel processing. Additionally, Intel’s real-time libraries ensure low-latency execution, with tools for performance tuning and EtherCAT-based industrial communication.

To ensure seamless integration with robotic hardware, the SDK runs on a real-time optimized Linux BSP. It includes support for optimized EtherCAT and camera drivers, along with Intel-specific features such as Speed Shift Technology and Cache Allocation to enhance power efficiency and performance. These system-level enhancements allow the software stack to deliver high responsiveness, making it suitable for real-time robotics applications.

Overall, the Embodied Intelligence SDK provides a highly optimized, AI-driven framework for robotics and Embodied Intelligence, combining computer vision, motion planning, real-time processing, and large-scale AI models into a cohesive system. By leveraging Intel’s hardware acceleration and software ecosystem, it enables next-generation robotic applications with enhanced intelligence, efficiency, and adaptability.

Release Note

Click each tab to learn about the new and updated features in each release of Intel® Embodied Intelligence SDK.

Embodied Intelligence SDK v25.36Embodied Intelligence SDK v25.15

Embodied Intelligence SDK v25.36 enhances model optimization capabilities with Intel® OpenVINO™ and provides typical workflows and examples, including Diffusion Policy (DP), Robotic Diffusion Transformer (RDT), Improved 3D Diffusion Policy (IDP3), Visual Servoing (CNS) and LLM Robotic Demo. This release also updated Real-time optimized BKC to improve on AI/Control performance, and supporting on Intel Arc B-Series Graphic card (B570).

New Features:

• Updated Real-time optimization BKC, including BIOS and runtime optimization, balancing performance with AI & Control consolidation.

• Added support for Intel Arc B-Series (Battlemage) Graphics card (B570).

• Fixed deadlock issue when reading i915 perf event in Preempt-RT kernel.

• New EtherCAT Master stack features supporting user-space EtherCAT Master and multiple EtherCAT masters.

• Added Diffusion Policy pipeline with OpenVINO optimization.

• Added Robotics Diffusion Transformer (RDT) pipeline with OpenVINO optimization.

• Added Improved 3D Diffusion Policy (IDP3) model with OpenVINO optimization.

• Added Visual Servoing (CNS) model with OpenVINO optimization.

• Provided new tutorials for typical AI model optimization with OpenVINO.

• ACRN initial enablement on ARL platform.

• Added new Dockerfile to build containerized RDT pipeline.

Known Issues and Limitations

• ACRN feature and performance

  1. iGPU performance degradation observed when using passthrough iGPU to VM on ACRN.

  2. Display becomes unresponsive in VMs when running concurrent AI workloads with iGPU SR-IOV enabled on ACRN.

The following model algorithms were added and optimized by Intel® OpenVINO™:

Algorithm	Description
Qwen2.5VL	Qwen2.5VL
Whisper	Whisper
FunASR (Automatic speech recognition)	Refer to the FunASR Setup in LLM Robotics sample pipeline
Visual Servoing - CNS	A graph neural network-based solution for image servo utilizing explicit keypoints correspondence obtained from any detector-based feature matching methods
Diffusion Policy	A visuomotor policy learning model in the field of robotic visuomotor policy learning, which represents policies as conditional denoising diffusion processes
Improved 3D Diffusion Policy (iDP3)	A diffusion policy model enhancing capabilities for 3D robotic manipulation tasks
Robotic Diffusion Transformer (RDT-1B)	A diffusion-based foundation model for robotic manipulation

The following pipelines were added:

Pipeline Name	Description
Diffusion Policy	An innovative method for generating robot actions by conceptualizing visuomotor policy learning as a conditional denoising diffusion process
Robotics Diffusion Transformer (RDT)	A RDT pipeline provided for evaluating the VLA model on the simulation task
LLM Robotics Demo	A code generation demo for robotics, interacting with a chatbot utilizing AI technologies such as large language models (Phi-4) and computer vision (SAM, CLIP)