Software Index

The table below lists the software solutions, sample pipelines, and tutorials available across this documentation. The Domains column categorizes each entry to help you find relevant material for your application.

Reference	Domains	Description
RealSense Camera with ROS 2 Sample Application	Sensors, Middleware	Integrates a RealSense camera with ROS 2 to stream color and depth data, launch camera nodes, and visualize the feed in RViz2.
3D Pointcloud Groundfloor Segmentation for RealSense Camera and 3D LiDAR	Sensors, AI	Intel algorithm that classifies 3D point clouds from RealSense or LiDAR sensors into ground, elevated surfaces, and obstacles for navigation over challenging terrain.
Multi-Camera Object Detection Powered by OpenVINO™	AI, OpenVINO™, Sensors	Runs OpenVINO™-optimized YOLOv8 object detection and segmentation in parallel across up to four USB or GMSL cameras.
OpenVINO™ Object Detection Tutorial	AI, OpenVINO™, Sensors, Middleware	Deploys a ROS 2 OpenVINO™ node for object detection with selectable CPU, GPU, or NPU inference devices.
OpenVINO™ Yolov8 Tutorial	AI, OpenVINO™, Sensors, Middleware	Installs a ROS 2 OpenVINO™ node and runs a YOLOv8 segmentation model on the CPU using a RealSense camera image as input.
OpenVINO™ Tutorial with Segmentation	AI, OpenVINO™, Sensors, Middleware	Runs a ROS 2 OpenVINO™ semantic segmentation model on CPU or GPU using a RealSense camera image as input.
Collaborative Visual SLAM	Autonomous Mobile Robot, SLAM	Multi-robot visual SLAM optimized with SSE/AVX2 instruction sets for map building and merging on Intel® CPUs and GPUs.
FastMapping Algorithm	Autonomous Mobile Robot, Sensors	Intel-optimized octomap implementation that builds 3D voxel maps from RealSense depth camera data for efficient environment representation.
ADBSCAN Follow-me	Autonomous Mobile Robot, AI, Sensors	Adaptive DBSCAN person detection and tracking from 2D/3D LiDAR or RealSense point clouds, with Gazebo simulation and real-robot deployment examples.
ITS Path Planner ROS 2 Navigation Plugin	Autonomous Mobile Robot, Navigation	Intel patented global path planner delivering 20-30x speedup over A* for the ROS 2 Navigation2 stack.
Robot Re-localization Package for ROS 2 Navigation	Autonomous Mobile Robot, Navigation	Re-localization algorithm that rapidly recovers robot pose in Nav2 after sensor glitches or environment disturbances.
GPU ORB Extractor	Autonomous Mobile Robot, SLAM	GPU-accelerated keypoint and descriptor extraction for Visual SLAM front-ends, with OpenCV and OpenCV-free APIs.
Deploy Robot Teleop Using a Keyboard	Autonomous Mobile Robot	Validates motor control on a deployed robot using keyboard teleoperation before running autonomous workloads.
Deploying wandering	Autonomous Mobile Robot, Navigation	Deploys the Wandering autonomous exploration pipeline on a physical robot using RTAB-Map and Nav2.
Simulated Robotics with Gazebo	Autonomous Mobile Robot, Simulation	Introduces simulating robots as digital twins in Gazebo to test robotics applications before real-world deployment.
Simulating wandering in Gazebo	Autonomous Mobile Robot, Simulation, Navigation	Simulates the full Wandering pipeline in Gazebo with mapping, frontier exploration, and Nav2-based navigation.
Gazebo Pick & Place Demo	Middleware, Manipulation, Simulation	Coordinates two UR5 arms and a TurtleBot3 AMR on a conveyor line using MoveIt2 and Nav2 in Gazebo Classic.
Vision and Controls Simulation Demo	Stationary Arm, Simulation, Manipulation	Simulates a UR5e arm, Robotiq gripper, and RealSense camera in Gazebo with MoveIt2 motion planning.
Vision and Controls Deployment Demo	Stationary Arm, Manipulation	Deploys a physical UR5e arm with 2D object detection, 3D pose estimation, and pick-and-place execution.
Imitation Learning - ACT	Humanoid, AI, OpenVINO™, Manipulation	Imitation learning pipeline using Action Chunking with Transformers, optimized with OpenVINO™, for fine manipulation in simulation and on real ALOHA robots.
Model Predictive Control Demo	Humanoid, AI, Manipulation	Combines ACT imitation learning with OCS2 model predictive control and MuJoCo simulation for perception-action manipulation control.
Diffusion Policy	Humanoid, AI, OpenVINO™, Manipulation	Visuomotor diffusion-policy pipeline for the Push-T manipulation task, with Transformer- and CNN-based variants optimized by OpenVINO™.
VSLAM: ORB-SLAM3	Humanoid, SLAM	Real-time feature-based Visual SLAM supporting monocular, stereo, and RGB-D cameras, with EUROC dataset and RealSense demos.
LIO SLAM: Point-LIO	Humanoid, SLAM	Robust high-bandwidth LiDAR-inertial odometry and mapping on Intel® platforms.
LIO SLAM: FAST-LIO2	Humanoid, SLAM	High-performance LiDAR-inertial odometry and mapping with FAST-LIO2 on Intel® platforms.
Fast-LIVO2 Demo	Humanoid, SLAM	Direct LiDAR-Inertial-Visual Odometry and mapping fusing LiDAR and camera streams.
LLM Robotics Demo	Humanoid, AI, Manipulation	Code-generation pipeline combining an LLM (Phi-4), vision models (SAM, CLIP), and a JAKA arm for voice- or text-commanded robot control.
Robotics Diffusion Transformer (RDT)	Humanoid, AI, OpenVINO™, Manipulation	Bimanual manipulation foundation model with a unified action space, running in MuJoCo simulation and on real ALOHA robots with OpenVINO™ optimization.
Pi0.5 with Real-Time Chunking	Humanoid, AI, OpenVINO™, Manipulation	Vision-Language-Action pipeline pairing a PaliGemma VLM with a flow-matching policy and real-time chunking for smooth high-frequency control, accelerated with OpenVINO™.
OpenClaw + AgenticROS Deployment	Humanoid, AI	Integrates OpenClaw and AgenticROS agent frameworks with OpenVINO™ Model Server for robot control in Gazebo.
GEAR-SONIC WBC	Humanoid, AI, Whole-Body Control	Whole-body control optimization of the SONIC WBC inference pipeline on Intel® Core™ Ultra platforms.
Action Chunking with Transformers - ACT	AI, OpenVINO, Manipulation	Imitation learning model that predicts action chunks with Transformers for fine manipulation, including conversion to OpenVINO™ IR.
Diffusion Policy (Model)	AI, OpenVINO, Manipulation	Visuomotor policy using conditional denoising diffusion to handle multimodal action distributions, with low-dim and image variants and OpenVINO™ conversion.
Robotics Diffusion Transformer (RDT-1B)	AI, OpenVINO, Manipulation	1.2B-parameter diffusion foundation model for manipulation pre-trained on 46 datasets, with OpenVINO™ IR conversion guidance.
General-purpose robot foundation model (Pi0)	AI, OpenVINO, Manipulation	Vision-Language-Action foundation model pairing a PaliGemma VLM with an action-expert diffusion transformer, including OpenVINO™ conversion.
BC-RNN & BC-Transformer	AI, OpenVINO, Manipulation	Behavior cloning models using RNN or Transformer backbones to map observations to actions from expert demonstrations.
Visual Servoing - CNS	AI, OpenVINO, Manipulation	Graph neural network image-based visual servo policy achieving sub-millimeter precision at real-time (~40 fps) rates.
GraspNet - Baseline	AI, OpenVINO, Manipulation	Grasp generation model trained on GraspNet-1Billion that predicts scored 6-DoF grasp poses from point clouds.
Feature Extraction Model: SuperPoint	AI, OpenVINO, SLAM	Self-supervised interest point detector and descriptor generator with homographic adaptation for cross-domain generalization.
Feature Tracking Model: LightGlue	AI, OpenVINO, SLAM	Lightweight transformer feature matcher with adaptive depth and width for efficient correspondence in 3D reconstruction and localization.
Improved 3D Diffusion Policy (iDP3)	AI, OpenVINO, Manipulation	Enhanced 3D manipulation policy that encodes point clouds with a 3D visual encoder and generates actions via diffusion.
Bird’s Eye View Perception: Fast-BEV	AI, OpenVINO, Navigation	Efficient multi-scale bird’s-eye-view perception model for obstacle avoidance, path planning, and spatial awareness.
Monocular Depth Estimation: Depth Anything V2	AI, OpenVINO, Sensors	Monocular depth estimation foundation model (25M-1.3B parameters) for cost-effective depth perception without LiDAR.
Dual-arm Manipulation Foundation Model: GR00T N1D7	AI, OpenVINO, Manipulation	Dual-arm manipulation foundation model optimized with OpenVINO™ for real-time inference.
Pi0.5 Model Optimization	AI, OpenVINO	Conversion, INT8 compression via Neural Network Compression Framework, latency/throughput benchmarking, and validation of the Pi0.5 VLA model.
Geti™	AI, Developer Tools	Interactive computer-vision platform for data labeling, training, optimizing, and exporting OpenVINO-ready vision models.
Action Chunking with Transformers - ACT	AI, OpenVINO™, Manipulation	Imitation learning model that predicts action chunks with Transformers for fine manipulation, including conversion to OpenVINO™ IR.
Diffusion Policy (Model)	AI, OpenVINO™, Manipulation	Visuomotor policy using conditional denoising diffusion to handle multimodal action distributions, with low-dim and image variants and OpenVINO™ conversion.
Robotics Diffusion Transformer (RDT-1B)	AI, OpenVINO™, Manipulation	1.2B-parameter diffusion foundation model for manipulation pre-trained on 46 datasets, with OpenVINO™ IR conversion guidance.
General-purpose robot foundation model (Pi0)	AI, OpenVINO™, Manipulation	Vision-Language-Action foundation model pairing a PaliGemma VLM with an action-expert diffusion transformer, including OpenVINO™ conversion.
BC-RNN & BC-Transformer	AI, OpenVINO™, Manipulation	Behavior cloning models using RNN or Transformer backbones to map observations to actions from expert demonstrations.
Visual Servoing - CNS	AI, OpenVINO™, Manipulation	Graph neural network image-based visual servo policy achieving sub-millimeter precision at real-time (~40 fps) rates.
GraspNet - Baseline	AI, OpenVINO™, Manipulation	Grasp generation model trained on GraspNet-1Billion that predicts scored 6-DoF grasp poses from point clouds.
Feature Extraction Model: SuperPoint	AI, OpenVINO™, SLAM	Self-supervised interest point detector and descriptor generator with homographic adaptation for cross-domain generalization.
Feature Tracking Model: LightGlue	AI, OpenVINO™, SLAM	Lightweight transformer feature matcher with adaptive depth and width for efficient correspondence in 3D reconstruction and localization.
Improved 3D Diffusion Policy (iDP3)	AI, OpenVINO™, Manipulation	Enhanced 3D manipulation policy that encodes point clouds with a 3D visual encoder and generates actions via diffusion.
Bird’s Eye View Perception: Fast-BEV	AI, OpenVINO™, Navigation	Efficient multi-scale bird’s-eye-view perception model for obstacle avoidance, path planning, and spatial awareness.
Monocular Depth Estimation: Depth Anything V2	AI, OpenVINO™, Sensors	Monocular depth estimation foundation model (25M-1.3B parameters) for cost-effective depth perception without LiDAR.
Dual-arm Manipulation Foundation Model: GR00T N1D7	AI, OpenVINO™, Manipulation	Dual-arm manipulation foundation model optimized with OpenVINO™ for real-time inference.
Pi0.5 Model Optimization	AI, OpenVINO™	Conversion, INT8 compression via NNCF, latency/throughput benchmarking, and validation of the Pi0.5 VLA model.
Geti™	AI, Developer Tools	Interactive computer-vision platform for data labeling, training, optimizing, and exporting OpenVINO™-ready vision models.
Wandering AMR Pipeline Benchmark	Benchmarking, Autonomous Mobile Robot	Automated benchmarking of the Wandering AMR pipeline, measuring latency, resource usage, and optional GPU/NPU KPIs across runs.
Pick & Place Pipeline Benchmark	Benchmarking, Manipulation	Automated benchmarking of the multi-robot Pick & Place simulation, capturing lifecycle metrics and aggregated KPIs.
RealSense2 Camera Benchmark	Benchmarking, Sensors	Automated benchmarking for USB and GMSL RealSense camera pipelines measuring FPS, latency, and system load.