Skip to main content

RDK X Series Release History

RDK X5​

Notes:

  • After updating the system, please use the rdk-miniboot-update command to update the NAND firmware to the latest version.
  • Official new images are released regularly, containing the latest feature optimizations and bug fixes. Users can either download and install the latest image or update the system online.

Version: 3.5.0​

Version Information​

  • System Version: RDKOS V3.5.0
  • Release Date: April 2026
  • Platform: RDK X5

Version Update Overview​

1. Comprehensive Upgrade of Algorithm Inference Architecture

  • Starting from version 3.5.0, Python algorithm inference, object detection, semantic segmentation, and other modules have been uniformly upgraded to the hbm_runtime interface. This interface, implemented using pybind11, provides efficient access to the libdnn C++ library and supports high-performance model loading and neural network inference.

2. Comprehensive Refactoring of Python Examples

  • Based on the new hbm_runtime interface, all examples in /app/pydev_demo/ have been completely rewritten to ensure interface consistency, optimal performance, and simpler usage.

3. Synchronized Updates to Official Documentation

  • The RDK documentation has been fully revised based on hbm_runtime and the new examples in /app/pydev_demo/, fully aligning with the latest version.

4. HDMI Timing Optimization

  • Optimized HDMI timing parameters, significantly improving multi-resolution compatibility and display stability.

5. GPU Desktop Experience Optimization

  • Fixed related bugs, resolving issues such as GPU mouse cursor residue and the inability to use the VLC tool.

6. Synchronization with the Latest SDK Version

  • Integrated all latest cumulative updates, optimizations, and stability enhancements from the upstream SDK.

7. Multimedia Example Fixes

  • Fixed known issues in /app/multimedia_samples to improve runtime stability.

8. API Enhancements

  • Added sp_open_camera_v3 to the C++ interface.
  • Added the open_cam method to the Python Camera object.
  • Supports configuring cropping parameters when opening the camera.

9. New Hardware Support

  • Added support for the Horned Sung Audio Subboard.
  • Added interface support for BMI08x and ICM42688 IMU sensors.

Version: 3.4.1​

Version Information​

  • System Version: RDKOS V3.4.1
  • Release Date: December 2025
  • Platform: RDK X5

Version Update Overview​

This RDK OS release is fully aligned and synchronized with the latest underlying platform software X5 SDK 1.1.1. Compared to the previous version (based on X5 SDK 1.0.17), this update integrates feature enhancements, stability fixes, and performance optimizations from the past six months. After adaptation, verification, and system tailoring, all core capabilities have been fully integrated into RDK OS. Building on the RDK Ubuntu 22.04 and TROS Robot Operating System experience, this version provides developers with more comprehensive platform support and a consistent runtime experience that matches the latest SDK.

(1) System and Boot Stability

  • EMMC/SD Boot Reliability Enhancement: Adapted for multiple brands of SD/TF cards, significantly improving success rates and compatibility for cold boots and hot restarts.
  • System Configuration Consistency Optimization: Improved srpi-config audio configuration flow, synchronizing with Ubuntu PulseAudio channel switching to reduce additional user configuration costs.

(2) Media Pipeline and Display Adaptation

  • Massive Media Pipeline Fixes: Covers stability enhancements and exception scenario fixes for modules including ISP, VIO, encoding/decoding, etc.
  • Sample Programs Cleanup and Completion: Updated interface calls, filled in missing examples, and improved usability.
  • HDMI Output Capability Enhancement: Improved multi-resolution support and compatibility, enhancing display stability.
  • LCD Full-Size Adaptation:
    • Added full adaptation for the 11.9-inch Waveshare ultra-wide screen.
    • Fixed touch coordinate anomalies on multiple screens.

(3) Peripheral Interfaces and Application Development Capabilities

  • WiringPi Compatibility Fixes: Improved ease of use for scripts and peripheral control.
  • hobot-gpio Functionality Expansion: Supports multi-channel PWM configuration.

(4) Documentation and Development Experience

  • Comprehensive Rewrite of "Chapter 3: Basic Application Development": Re-planned chapter structure, supplemented key concept explanations, enriched sample code and practical operation demonstrations, significantly improving document readability and learning path coherence.
  • Fixed and Supplemented Over 20 Documentation Issues: Covers incomplete interface descriptions, missing examples, inconsistent configuration steps, etc., making the overall documentation more accurate, standardized, and easier to use.

(5) TogetheROS

  • Fixed Image Processing Acceleration Module: Fixed a bug in hobot_cv using vse to accelerate image resize.
  • Fixed Onboard Model Inference Framework: Fixed an issue in hobot_dnn where inference latency statistics were incorrect.
  • Fixed MIPI Cam and Algorithm Modules: Fixed an issue where incorrect installation paths for ros component shared objects (so) caused runtime loading failures.
  • Optimized Image Codec Module: Optimized hobot_codec configuration, removed invalid configuration parameters, and added configuration parameters for debugging.

Version: 3.3.3​

Version Information​

  • System Version: RDKOS V3.3.3
  • Release Date: October 2025
  • Platform: RDK X5

Version Update Overview​

(1) System and Driver Updates

  • Added Yuguang-SC132GS stereo camera module driver
  • srpi-config functionality enhancement: Supports automatic interface pin multiplexing, added MIPI screen selection
  • GPU desktop stability improvements
  • Wi-Fi driver upgraded to 2025_0410, improving wireless connection stability
  • Flash connection interface enhancements:
    • Supports usage on macOS systems
    • Added RDK Studio USB connection functionality
  • Multiple bug fixes and performance optimizations

(2) Documentation Optimizations

  • V4L2 usage documentation optimized, added examples and descriptions
  • Display screen usage documentation optimized, improved typical screen configuration descriptions
  • Other documentation detail optimizations and structural adjustments

(3) Ecosystem Support

The following ecosystem libraries have been released to PyPI:

Library NameVersionUpdate Content
hobot-dnn-rdkx53.0.6Added dependency numpy >= 1.26.4
hobot-vio-rdkx53.0.6Fixed cropping not taking effect; sp_open_camera interface default output size is 1920×1080

TROS Update Notes (V2.4.3)​

  • Monocular MIPI image acquisition: Supports starting multiple image acquisition channels
  • Stereo MIPI image acquisition: Added support for SC132GS stereo camera

How to Obtain​

RDK X5 V3.3.3 version has been synchronized to:

  • D-Robotics Developer Community Download Center, download and upgrade
  • sudo apt update && sudo apt upgrade; direct local upgrade

TROS Algorithm Upgrade & Version: 3.2.3​

Image Updates​

  • 20250610
    • Updated ROS repository GPG signing key
    • Supports real-time kernel switching
    • Integrated the latest miniboot firmware; after burning NAND with rdk-miniboot-update, more memory is released to the system
  • 20250604
    • Fixed system boot failure issue on Hiwonder carrier board

System Layer Updates​

  • Desktop Display Optimization: Ubuntu desktop now supports 3D GPU accelerated rendering for a smoother visual experience
  • Audio Sub-board Adaptation Added: Supports Waveshare WM8960 Audio HAT and Hiwonder carrier board, accelerating voice solution integration
  • WIFI Driver Upgrade: Enhanced connection stability in weak signal environments, optimized auto-reconnection during sleep/wake
  • Sensor Acquisition Framework Expansion: Integrated V4L2 framework, adapted for imx477, ov5647, imx219; more sensor support coming soon
  • Network & Remote Optimization: Default switched to iptables legacy mode, improved VNC display smoothness, remote desktop is lag-free
  • Interface Enhancement: CAN interface stability optimized, solved high-speed data packet loss; srpi-config tool adds Uart7 support for better serial expansion
  • Storage Compatibility Optimization: Improved SD card compatibility, supports more storage card models

Application Layer Updates​

  • Voice Capability Enhancement: Added ASR voice recognition solution for more efficient voice algorithm development
  • Stereo Depth Algorithm Upgrade: Optimized depth estimation algorithm, significantly improved detection speed and accuracy
  • Multimodal Demo Integration: Built-in edge large model multimodal demos, enabling custom application solutions in 3 minutes

Stereo Algorithm Upgrade & Version: 3.1.1​

Core Features​

  • System Backup: Brand new rdk-backup tool for one-click system backup and easy image generation. See rdk-backup introduction
  • Configuration Management: Supports using config.txt to configure 40pin pin initialization state during U-Boot, improving system boot stability
  • Touchscreen Enhancement: Added double-tap and long-press actions; long-press simulates right-click for more flexible screen control
  • Device Tree Overlay Support: Added 1_wire device tree overlay (dtoverlay) example for more hardware customization options
  • Stereo Algorithm Upgrade: StereoNet depth algorithm upgraded for significantly improved depth results; added ZED camera support for stereo image acquisition, easily build intelligent vision systems with StereoNet
  • New Application: Smart Video Box: hobot_rtsp_client supports RTSP streaming, decoding, intelligent inference, and web-based result display for quick edge AI integration
  • Open Vocabulary Detection: DOSOD: hobot_dosod provides an open-vocabulary detection algorithm developed by D-Robotics, enabling flexible voice interaction on edge devices

Bug Fixes & Optimizations​

  • Fixed efficientnasnet_m_300x300_nv12.bin model import issue
  • Optimized SD card protocol support for better compatibility
  • Fixed black screen issue caused by portrait display
  • dnn_node fixed YOLOv8-seg post-processing box out-of-bounds crash
  • hobot_codec fixed frame rate calculation error
  • hobot_stereonet_utils removed non-startable launch files
  • Resolved multi-channel I2C detection issues, added LPWM switch configuration

Version: 3.1.0​

New Features​

  • Added key sleep and wake-up function
  • Enabled 40Pin secondary functions

Optimizations​

  • Bug fixes
  • Corrected CAN frequency anomaly
  • Supports more sensors and resolutions

Notes:
When upgrading from an older version to this version using apt update && apt upgrade, you must first uninstall tros-humble-stereonet-model and then install the tros-humble-hobot-stereonet package.

sudo apt-get remove tros-humble-stereonet-model
sudo dpkg --remove --force-all tros-humble-stereonet-model
sudo apt install -y tros-humble-hobot-stereonet

Version: 3.0.1​

New Features​

  • Provided Server version firmware
  • Supported 7 Waveshare MIPI DSI LCD screens for desktop display and touch
  • Supported running sample programs as non-root users

Optimizations​

  • Bug fixes
  • Supports more sensors and resolutions
  • Improved high-resolution display stability

Version: 3.0.0​

First release of RDK X5 firmware, based on Ubuntu 22.04, providing rich multimedia and algorithm samples, supporting robot application development for various scenarios.

RDK X3​

Version: 3.0.0​

New Features​

  • Supports Ubuntu 22.04

Version: 2.1.0​

New Features​

  • Improved srpi-config system configuration tool: supports Wi-Fi connection, SSH/VNC toggle, enabling/disabling peripheral buses on 40pin, localization, CPU overclocking, ION memory size, etc.
  • Supports /boot/config.txt system config file: dtoverlay, CPU overclock, IO boot state, etc.
  • Added yolov5s v6/v7 model samples

Optimizations​

  • HDMI display now outputs boot logs and user shell for easier use
  • Supports more HDMI resolutions for better compatibility
  • Optimized pre-installed software for Desktop/Server, removed redundancy, added essentials like VLC
  • Simplified Desktop menu bar
  • Bluetooth enabled by default
  • Added C++ interface for post-processing to improve efficiency
  • Used udisk2 for auto-mounting USB drives, fixed NTFS access issues
  • Users can retain VNC password file
  • VNC service is off by default to save resources; can be enabled via srpi-config
  • RDK X3 v2.1 and RDK Module boards run up to 1.5GHz (normal), 1.8GHz (overclocked)

Bug Fixes​

  • Removed redundant Wi-Fi driver kernel logs
  • Changed apt source domain to archive.d-robotics.cc

Other Updates​

  • Supports Chromium browser (sudo apt install chromium)

Version: 2.0.0​

This release brings many anticipated features and improvements for a better development experience and broader application support.

Open Source​

  • Fully open source OS, including core and functional modules. Developers can freely view and modify for customization and optimization
  • Detailed code documentation and comments to help developers understand and use
  • Community contributions and discussions welcome; source code maintained at D-Robotics

RDK X3 Module Support​

  • Introduced new core board developer kit: RDK X3 Module
  • Smaller size, compatible with Raspberry Pi CM4 interface
  • Developers can choose third-party carrier boards as needed to expand functions and applications

Other Updates​

  • Optimized existing features, fixed known issues and vulnerabilities, improved OS stability and performance
  • Revised documentation for more comprehensive and accurate technical materials and guides
  • Provided lower-level APIs for secondary development and integration

Version: 1.0.0​

  1. cat /etc/version shows 1.x.x
  2. rdkos_info command not available
  3. Old system, no open source code, uses old manual, minor bugs fixed in new images
  4. D-Robotics source domain and key changed

TogetheROS.Bot​

tros-humble​

Version: 2.4.0 (2025-05-12)​

New Features:

  • Supports RDK S100 platform.

Version: 2.3.3 (2025-04-30)​

New Features:

Version: 2.3.2 (2025-01-15)​

Feature Changes:

New Features:

Bug Fixes:

Version: 2.3.1 (2024-11-20)​

Feature Changes:

  • Upgraded opencv dependency from 3.4.5 to 4.5.4 (latest Ubuntu 22.04 release).

New Features:

Bug Fixes:

Version: 2.3.0 (2024-09-19)​

New Features:

Version: 2.2.0 (2024-04-11)​

Feature Changes:

  • Based on TROS Foxy 2.1.3, adapted for Ubuntu 22.04 and ROS2 Humble.
  • TROS install path changed to /opt/tros/humble.
  • No longer provides tros-ros-base package, directly depends on standard ROS2 release.
  • Uses ROS2 fastdds zero-copy communication, QoS Reliability set to BEST_EFFORT.
  • hobot_dnn, hobot_audio refactored based on libdnn.
  • hobot_trigger adapted for ROS2 Humble rosbag2.

New Features:

  • robot_dev_config adds bloom build script.
  • hobot_mipi_cam node adds frame_ts_type config.
  • Added hobot_shm node for ROS2 zero-copy environment config.

Bug Fixes:

  • Fixed compiler upgrade compatibility issues.
  • Fixed board-side build ROS2 pkg path dependencies.

tros-foxy​

Version: 2.1.3 (2024-03-11)​

Feature Changes:

  • JPEG compressed image type changed from sensor_msgs::msg::Image to sensor_msgs::msg::CompressedImage, supports foxglove/ros2 rqt tools.
  • Unified jpeg/mjpeg config, removed jpeg-compressed/mjpeg-compressed configs.
  • Added TROS_DISTRO env variable, config path changed to /opt/tros/${TROS_DISTRO}/lib.

Version: 2.1.2 (2024-01-19)​

New Features:

  • Refactored hobot_usb_cam, supports more format configs and transcoding.
  • hobot_audio updated voice SDK, supports 2mic/4mic boards, adds micphone_name config.

Bug Fixes:

  • hobot_rgbd_cam node fixed step field error.
  • hobot_tts fixed audio playback failure.
  • hobot_llm removed config device tree file, updated README, supports ION memory config via command tool.

Version: 2.1.1 (2023-11-03)​

New Features:

  • Added hobot_chatbot node for on-board voice chat.

Bug Fixes:

  • hobot_tts node fixed exit issue caused by special characters.

Version: 2.1.0 (2023-09-14)​

Feature Changes:

  • tros-ros-base upgraded to latest ROS2 foxy source.
  • Using ROS2 foxy only requires source /opt/tros/setup.bash.

New Features:

  • hobot_tts node adds audio device parameter.
  • Added hobot_llm node for edge LLM.
  • hobot_codec node adds jpeg-compressed config.

Bug Fixes:

  • hobot_mipi_cam node fixed RGB format step field error.

Version: 2.0.2 (2023-08-28)​

Feature Changes:

  • ROS2 source switched to Tsinghua mirror.

New Features:

  • tros.b script adds permission check, auto switches to root.
  • hobot_audio node adds audio device number parameter.
  • hobot_trigger node adds std_msg topic trigger.

Bug Fixes:

  • hobot_cv node fixed crop&resize simultaneous processing failure.
  • hobot_mipi_cam node startup error log issue.
  • hobot_visualization node launch config invalid issue.

Version: 2.0-Release (2.0.1) (2023-06-10)​

Feature Changes:

  • Upgraded voice algorithm, improved ASR.
  • Algorithm sample model_name config auto parsing.
  • tros.b package no longer includes nav2, use apt to install latest nav2.

New Features:

  • Supports RDK Ultra platform.
  • Added Trigger event, rosbag visualization node.
  • USB image acquisition node supports device number auto-adaptation.
  • Added VIO algorithm node.
  • Added hobot_tts node.
  • Added hobot_centerpoint, hobot_bev, hobot_stereonet node.

Bug Fixes:

  • Upgraded RDK X3 easydnn/dnn, fixed operator crash.
  • Fixed RGBD image acquisition node depth data error.

Other Updates:

  • Optimized human detection/tracking node for resolution adaptation.
  • Fixed orb_slam3 build script path error.

Version: 2.0-Beta (2.0.0) (2023-05-29)​

2.0-Beta (2.0.0) is the first 2.x version, 1.x users are recommended to upgrade.

Feature Changes:

  • Code hosting migrated to GitHub.
  • Integrated efficient package management.

New Features:

  • Supports RDK X3 Module.
  • hobot_audio adds ASR result output.

Bug Fixes:

  • Fixed dnn_node MobileNet_SSD multi-thread crash.
  • Fixed X86 dnn_node, hobot_codec, hobot_image_publisher build failures.

Other Updates:

  • Updated sample launch scripts.
  • Websocket display updated D-Robotics logo.

TogetheROS-V1.x​

V1.1.6​

  • Added X86 platform support, including Hobot_DNN, Hobot_codec, etc.
  • Fixed audio codec driver auto-load I2C conflict.
  • Fixed USB camera video device node change issue.

V1.1.5​

  • Added hobot_audio support for linear/ring mic config.
  • Optimized hobot_sensors MIPI Camera timestamp accuracy.
  • Added hobot_image_publisher h265 encoded test video.
  • Fixed TROS package path dependency.
  • Fixed user file overwrite after system update.

V1.1.4​

  • Optimized hobot_sensors for more modules.
  • Added VIO raw/yuv image interface.
  • Added hobot_image_publisher MP4/H264/H265 video publishing.
  • Fixed hobot_dnn post-processing no return value.
  • Added hobot_cv benchmark tool.
  • Optimized toolchain module, added bug fix guidance.
  • Fixed USB3.0 and HDMI capture card compatibility.

V1.1.3​

  • Optimized toolchain, one-click model quantization.
  • Added C RTSP streaming decode sample.
  • Added hobot_cv image padding and nv12 processing.
  • Optimized hobot_codec log output.
  • Optimized hobot_sensor, added more camera support.

V1.1.2​

  • Added on-board docker support.
  • Added 2D garbage detection sample.
  • Added C/C++ multimedia and algorithm dev interfaces.
  • Added dnn_node_sample package usage sample.
  • Added adaptive multi-core model inference.
  • Added perception result confidence info.

V1.1.1​

  • Optimized vision/image module error prompts.
  • Optimized smart voice algorithm samples.
  • Added DOA sound source localization sample.
  • Integrated Navigation2 source.
  • User manual adds AI toolchain guide.

V1.1.0​

  • Opened kernel driver build environment.
  • Optimized TF card compatibility.
  • Added car line-following, auto-parking samples.
  • Added more camera adaptations.
  • Added Sunny Optical RGBD module adaptation.
  • hobot_codec fixed image crash.
  • Supports output frame rate control.
  • Added EDID detection, adaptive resolution.
  • Added joystick remote control support.
  • Added xfs/ntfs filesystem support.

V1.0.5​

  • Added parking scenario algorithm sample.
  • Added local image publishing tool.
  • Added hobot_cv neon accelerated filtering.
  • hobot_dnn fixed inference output frame rate error.
  • Supports camera type selection for human detection, etc.
  • hobot_codec fixed image step parameter error.
  • System software adds restart_network command.

V1.0.4​

  • Added ORB-SLAM3 and optimization.
  • hobot_dnn fixed inference errors, upgraded prediction library.
  • hobot_cv supports pyramid output.
  • System software adds RTSP video hardware decoding, Python API.
  • Fixed AI model I2C error, python pwm period setting, spi write/read issues.
  • Fixed Ethernet and wifi simultaneous startup error.
  • Added 4-mic audio sub-board instructions.

V1.0.3​

  • Added ROS2 vision_opencv adaptation.
  • hobot_audio optimized voice recognition, changed default wake word.
  • hobot_cv added crop/resize/rotate interfaces.
  • hobot_sensor optimized NV12 to RGB efficiency.
  • hobot_msgs optimized keypoint info, added audio_msg.

V1.0.2​

  • Added GC4663 mipi camera, RealSense camera adaptation.
  • Optimized AI samples, supports usb camera multi-resolution.
  • Optimized first boot time.
  • Fixed Desktop black screen issue.
  • mono2d_body_detection added keypoint confidence.
  • audio_control added voice control for robot movement.
  • hobot_audio added voice recognition.
  • hobot_websocket optimized web display.
  • hobot_msgs optimized keypoint info, added audio_msg.

V1.0.1​

  • Updated hobot-arm64-modules, fixed Desktop mouse display.
  • HobotDNN optimized async inference.
  • HobotCV optimized log output.
  • HobotSensor fixed rgbd sensor message publishing.
  • HobotWebsocket fixed memory leak.
  • HobotCodec optimized codec CPU usage.
  • Added HobotHDMI module.
  • mono2d_body_detection optimized performance stats and log.
  • hand_lmk_detection fixed memory leak and inference performance.
  • hand_gesture_detection fixed multi-hand detection and optimized performance.
Last updated on