{"product_id":"hiwonder-armpi-ultra-robot-arm-standard-version-with-depth-camera-raspberry-pi-5-8gb-amp-sd-card","title":"Hiwonder ArmPi Ultra Robot Arm Standard Version with Depth Camera, Raspberry Pi 5 (8GB) and SD Card","description":"\u003cp\u003e\u003cb\u003eFriendly Reminder:\u003c\/b\u003e\u003cbr\u003e* The ArmPi Ultra robotic arm is developed based on the Raspberry Pi 5 controller and uses the ROS2 operating system.\u003cbr\u003e* The advanced version supports embodied AI functions with AI large models and AI voice interaction, highly recommended for purchase!\u003cbr\u003e* The following content is for reference only. Specific configurations and functions depend on the version you choose. Thank you for your understanding!\u003cbr\u003e\u003cbr\u003e\u003cb\u003eVersion Comparison:\u003c\/b\u003e\u003cbr\u003e1. The starter version: with a monocular camera\u003cbr\u003e2. The standard version: with a depth camera\u003cbr\u003e3. The advanced version: with a depth camera and AI voice interaction box; supports AI large models\u003cbr\u003e\u003cbr\u003e\u003cb\u003eProduct Introduction:\u003c\/b\u003e\u003cbr\u003eArmPi Ultra is a desktop-level AI 3D vision robotic arm developed by Hiwonder for ROS educational scenarios. It is equipped with a 3D depth camera (or a monocular camera), combining 3D vision technology with robotic arm control. The system features six high-torque intelligent bus servos, a Raspberry Pi 5 main controller (optional), a multifunction expansion board, and an AI voice interaction box (optional). With built-in advanced inverse kinematics algorithms, it not only enables 3D motion control but also accurately recognizes, tracks, and grasps target objects in three-dimensional space.\u003cbr\u003e\u003cbr\u003eArmPi Ultra deploys multi-modal AI large models (only for the advanced version), integrating 3D vision with voice or text input, allowing it to deeply understand the environment, plan actions, and execute tasks flexibly. Using the Raspberry Pi's CPU, it can train machine learning models and easily implement AI projects such as waste sorting, object recognition, facial recognition, and human feature detection.\u003cbr\u003e\u003cbr\u003eArmPi Ultra features a compact modular design, supporting multi-modal expansion. When combined with Mecanum wheels, it can transform into a mobile handling robot, enabling mobile grasping and intelligent handling. It also supports motorized slide rail and conveyor belt expansions, allowing users to freely build dual-arm collaboration, item sorting, and other automated projects, making it suitable for diverse application scenarios.\u003cbr\u003e\u003cbr\u003e\u003cb\u003eFeatures:\u003c\/b\u003e\u003cbr\u003eArmPi Ultra AI Large Model ROS2 Robotic Arm\u003cbr\u003e* 3D Vision Depth Point Cloud (only applicable for versions with a depth camera)\u003cbr\u003e* Inverse Kinematics Hand-Eye Integration (only applicable for versions with a depth camera)\u003cbr\u003e* 3D Spatial Grasping (only applicable for versions with a depth camera)\u003cbr\u003e* Optional Embodied Intelligent Voice Interaction\u003cbr\u003e* High-Performance Hardware Configuration: Raspberry Pi 5 \u0026amp; STM32 controller, 25KG self-feedback bus servos, AI large model voice interaction box (only for the advanced version), optional 3D structured-light depth camera, multi-modal peripheral expansion\u003cbr\u003e* ROS2 Features: 3D spatial recognition, tracking \u0026amp; grasping, YOLO V8 depth vision image processing, MoveIt dynamic simulation, MediaPipe human–machine interaction, inverse kinematics hand-eye integration\u003cbr\u003e* Multi-Dimensional AI Large Model: Deploys text\/vision\/voice AI large models, AI large model voice interaction, multi-modal embodied intelligence applications, supports DeepSeek, Tongyi Qwen, and 01.AI\u003cbr\u003e\u003cbr\u003eA Professional Educational-Grade Robotic Arm Designed Specifically for Embedded Development Enthusiasts!\u003cbr\u003e* High-Performance Hardware Architecture: The ArmPi Ultra control system adopts a dual-brain architecture for efficient collaboration. It is equipped with self-feedback intelligent bus servos, a 3D depth camera, WonderEcho Pro voice interaction box, and other high-performance hardware, bringing ArmPi Ultra's capabilities to a new level! STM32 main controller + optional Raspberry Pi mainboard.\u003cbr\u003e* Multi-Modal AI Large Model: ArmPi Ultra deeply integrates text, voice, and vision AI large models, enabling natural and smooth human-like dialogue and control. Combined with vision and sensors, it can perform advanced embodied intelligence functions such as intelligent handling. It supports DeepSeek, 01.AI, Tongyi Qwen, and StepFun.\u003cbr\u003e* Advanced Inverse Kinematics Algorithm: ArmPi Ultra uses Hiwonder's self-developed inverse kinematics algorithm. Users can flexibly and accurately control robot movements by calling inverse kinematics functions. Integrated AI vision and AI large model algorithms empower personalized AI robotics innovation, providing a full-stack ROS development platform.\u003cbr\u003e* Multi-Modal Intelligent Expansion: ArmPi Ultra supports multi-modal expansions. When it is used with Mecanum wheel chassis, motorized slide rails, and conveyor belts, it can achieve mobile grasping, intelligent handling, and automation scenarios. Rich content design satisfies all your robotic arm imagination!\u003cbr\u003e\u003cbr\u003ePioneering 3D Vision Robotic Arms in Education\u003cbr\u003e* ArmPi Ultra is equipped with a 3D structured-light depth camera, integrating RGB and depth information fusion technology. It can accurately obtain depth data of objects, including color, shape, size, and position. Combined with advanced inverse kinematics algorithms, the robotic arm can lock onto, grasp, and sort targets in real time, enabling efficient intelligent workflows.\u003cbr\u003e* 3D Spatial Recognition: The robotic arm acquires depth data of the environment and target objects within its field of view.\u003cbr\u003e* 3D Spatial Tracking: The robotic arm can lock onto targets within its view and track them in real time using advanced inverse kinematics algorithms.\u003cbr\u003e* 3D Spatial Grasping: Thanks to advanced inverse kinematics algorithms, the robotic arm can precisely grasp target objects within its field of view.\u003cbr\u003e* 3D Spatial Sorting: The robotic arm can sort grasped targets into designated areas according to user commands.\u003cbr\u003e\u003cbr\u003eIndustry-First Integration of Multi-Dimensional AI Large Models with 3D Vision\u003cbr\u003e* ArmPi Ultra combines traditional vision-based sorting with multi-dimensional AI large models, including text, voice, and vision. This allows the robotic arm to understand user voice commands, deeply analyze the environment and objects within its field of view, and flexibly execute actions, significantly enhancing its intelligence and creating a truly AI-powered robotic arm.\u003cbr\u003e* AI Large Model 3D Sorting: The robotic arm can interpret user voice commands through a language model and deeply understand objects and the environment within its view through a vision model. Based on different voice instructions, it can autonomously grasp and transport target objects, sorting them into designated areas.\u003cbr\u003e* Semantic Understanding: The robotic arm uses large language models to reason about user voice commands and understand their meaning at a deeper level.\u003cbr\u003e* Environmental Perception: Through analysis by the vision model, the robotic arm can gain a deep understanding of objects and spatial layouts in the area.\u003cbr\u003e* Scene Understanding: The vision model enables the robotic arm to interpret the semantics of the environment, objects, spatial layout, and events within its field of view.\u003cbr\u003e* 3D Sorting: Thanks to advanced inverse kinematics algorithms, the robotic arm can precisely grasp target objects in view and perform grasping and sorting tasks according to user instructions.\u003cbr\u003e\u003cbr\u003eDepth Vision Enables Unified Hand–Eye Coordination\u003cbr\u003e* The end of the ArmPi Ultra robotic arm is equipped with a high-performance structured-light depth camera, allowing the system to work in three-dimensional space and determine an object's RGB, spatial coordinates, and depth information. This enables the measurement of distance, shape, height, volume, and more—substantially enhancing the robotic arm's autonomy and intelligence, and enabling richer AI project demonstrations.\u003cbr\u003e* 3D Depth Point Cloud Recognition: Through the depth camera's API, ArmPi Ultra can obtain depth maps, color images, and point cloud maps of the environment. It can then extract RGB data, spatial coordinates, and depth information of target objects, enabling shape recognition, object sorting, height measurement, material detection, and various AI demonstrations.\u003cbr\u003e* Depth-Based Distance Estimation: By acquiring depth point cloud data of objects, the system can determine the distance between the object and the depth camera, enabling object positioning, sorting, and tracking in three-dimensional space.\u003cbr\u003e* Height Measurement of Targets in a Region: By analyzing depth point cloud data, the system can determine the height of objects, allowing implementations such as detecting and removing items with abnormal height.\u003cbr\u003e\u003cbr\u003eAdvanced Inverse Kinematics Algorithm\u003cbr\u003e* ArmPi Ultra features Hiwonder's fully self-developed advanced inverse kinematics algorithm. Within the reachable 3D space, the end effector can move to any coordinate. With Python programming, path planning can also be achieved.\u003cbr\u003e* Target Detection and Joint Adaptive Adjustment: ArmPi Ultra can detect target objects within the recognition area, calculate their position coordinates and placement angles, and adaptively adjust each joint angle using the inverse kinematics algorithm to achieve free grabbing. Most other vision robotic arms on the market rely on fixed action groups for point grabbing and cannot perform 2D area detection or free grabbing.\u003cbr\u003e* 3D Space Motion Control: ArmPi Ultra uses inverse kinematics to achieve linear motion and path planning in 3D space, bringing motion control to a new level. This helps users understand interpolation principles and master the essence of inverse kinematics. Official detailed courses are provided.\u003cbr\u003e* DH Model and Inverse Kinematics Function Interface: Provides the ArmPi Ultra DH coordinate model, inverse kinematics analysis, and function interface. By inputting the end-effector coordinates, the inverse kinematics computation outputs each servo angle, making debugging easier and reducing project development time. Pure practical content to help you better understand robotic arm control. Official IK function interfaces are included.\u003cbr\u003e\u003cbr\u003eRGB+D Fusion Detection for 3D Free Grabbing\u003cbr\u003e* ArmPi Ultra's 3D depth camera fuses RGB information with depth data, enabling it to sense both object color and point-cloud depth, greatly enriching the geometric representation of spatial information. With the inverse kinematics algorithm, ArmPi Ultra can perform free grabbing, sorting, and handling in 3D space for advanced AI projects.\u003cbr\u003e* RGB+D 3D Space Recognition \u0026amp; Grabbing\u003cbr\u003e* Voice-Controlled 3D Grabbing and Sorting\u003cbr\u003e\u003cbr\u003eAI Vision Recognition and Target Tracking\u003cbr\u003e* Equipped with an RGB lens, ArmPi Ultra's 3D depth camera supports AI-powered image processing, enabling functions such as color recognition and tag identification.\u003cbr\u003e* Color Block Sorting: ArmPi Ultra can recognize and sort wooden blocks of different colors. Beyond standard colors, it also supports recognizing various custom colors.\u003cbr\u003e* Tag Recognition \u0026amp; Intelligent Stacking: ArmPi Ultra can identify different AprilTag codes, determine each tag block's position, and perform intelligent stacking.\u003cbr\u003e* Target Tracking: ArmPi Ultra can locate and track targets. With machine learning, it can also track more trained target objects.\u003cbr\u003e\u003cbr\u003eDeploy Three Types of AI Models for Full Embodied Intelligence\u003cbr\u003e* ArmPi Ultra integrates multi-modal AI models and connects online through the Qwen API to run large-model functions. It can also access other models via Alibaba Cloud, such as DeepSeek and 01.AI. ArmPi Ultra comes with an AI voice interaction box, greatly improving natural and fluent human–machine interaction. Combined with vision and sensors, it gains powerful perception, reasoning, and action capabilities, enabling more advanced embodied intelligence applications and creating a natural, flexible interaction experience.\u003cbr\u003e* The AI voice interaction box features a built-in noise-reduction microphone and a high-quality speaker. It is an intelligent voice module that integrates speech playback and recognition.\u003cbr\u003e* Large Language Model: ArmPi Ultra connects to Qwen in real time, enabling accurate text generation, translation, Q\u0026amp;A, summarization, and more. With a text-based large model, ArmPi Ultra functions like it has a super brain—able to understand your instructions and respond intelligently.\u003cbr\u003e* Speech Model: With the AI voice interaction box, ArmPi Ultra gains ears and a voice. Using advanced streaming end-to-end speech–language unified modeling and natural language processing, it can listen to every sentence you speak in real time and respond with highly natural, human-like speech.\u003cbr\u003e* Vision Model: The vision model acts as ArmPi Ultra's eyes, enabling it to understand and analyze image content. It can accurately recognize objects and provide detailed descriptions, including object names and features, achieving precise image recognition and intelligent visual interaction.\u003cbr\u003e\u003cbr\u003eMulti-modal Large-Model Applications\u003cbr\u003e* ArmPi Ultra integrates multi-modal large models, giving it cross-modal perception and understanding capabilities. It can flexibly switch and combine text, speech, and image models, making its interaction methods richer and smarter, and providing users with a highly creative interactive experience.\u003cbr\u003e* Voice Control: With semantic understanding powered by the large language model, ArmPi Ultra can interpret user commands and execute corresponding actions, enabling smooth and natural voice control.\u003cbr\u003e* Color Block Sorting: Through the large language model, ArmPi Ultra can recognize objects of different colors in the image and execute the corresponding actions based on user instructions to accurately complete color sorting.\u003cbr\u003e* Trash Sorting: Using the large language model, ArmPi Ultra can identify and understand different objects in the scene and follow user commands to precisely complete waste classification.\u003cbr\u003e* 3D Object Sorting: By combining the large language model with the depth camera, ArmPi Ultra can measure object volume within a target area and output corresponding actions according to user instructions, enabling 3D object sorting.\u003cbr\u003e\u003cbr\u003eEmbodied Intelligence Applications\u003cbr\u003e* With its high-performance hardware and deeply integrated multi-modal large models, ArmPi Ultra can perceive its environment, understand context, plan actions, and execute tasks flexibly. This enables advanced human–robot collaboration and intelligent task handling, delivering a more natural and efficient interaction experience.\u003cbr\u003e* Real-Time Detection: Using the vision model, ArmPi Ultra can analyze and understand image content and features within its field of view, then provide text or voice feedback through the large language model.\u003cbr\u003e* Visual Tracking: Through vision-model analysis, ArmPi Ultra can track any object within its visual range and lock onto it. Using inverse kinematics, it can continuously follow the target in real time.\u003cbr\u003e* Intelligent Handling: With the vision model, ArmPi Ultra can identify target objects in the scene, obtain their positional parameters, and adjust its posture in real time to grasp and place them at the designated position.\u003cbr\u003e* Intelligent Assistant: ArmPi Ultra can send the current visual frame to the vision model and, combined with user instructions, output the corresponding speech and actions.\u003cbr\u003e\u003cbr\u003eMulti-modal Expansion\u003cbr\u003e* ArmPi Ultra features a modular architecture that supports extensions such as a Mecanum-wheel chassis, electric slider, and electric conveyor belt. These expansions enable mobile handling, intelligent transportation, line-following movement, and advanced AI automation scenarios such as smart sorting systems.\u003cbr\u003e* Mecanum-Wheel Chassis Expansion: ArmPi Ultra supports a mobile chassis, allowing the robotic arm to perform mobile grabbing and intelligent handling. The chassis uses DC gear motors with encoders and can be freely detached, enabling ArmPi Ultra to function both as a desktop robotic arm and as a mobile robot.\u003cbr\u003e* Electric Slider Expansion: The electric slider is driven by a stepper motor and offers a maximum effective travel of 500mm, significantly extending the robotic arm's working range on the desktop and enabling more advanced AI creative applications.\u003cbr\u003e* Electric Conveyor Belt Expansion: The electric conveyor belt is equipped with a DC gear motor, PVC conveyor belt, and alloy frame. It provides smooth transportation of various objects, supports adjustable direction and speed, requires no additional power supply. And it is plug-and-play—perfect for AI-driven automated transport scenarios.\u003cbr\u003e\u003cbr\u003eDeep Learning and Model Training\u003cbr\u003e* ArmPi Ultra uses neural networks such as GoogLeNet, YOLO, and MTCNN, enabling it to perform deep learning on targets and generate trained models for more creative AI applications.\u003cbr\u003e* Upgraded YOLO V8 Algorithm: The robot adopts the YOLO v8 object detection algorithm, achieving millisecond-level real-time inference. It fully supports custom training and model quantization deployment, allowing cross-scene transfer learning and providing efficient, stable recognition for various AI innovation scenarios. It is highly compatible with mainstream embedded robotics platforms.\u003cbr\u003e* Improvements include: enhanced backbone structure, overall accuracy gains of +3.2% to +9.3%, inference acceleration of ~18%–30%, unified multitask architecture, and stronger scalability.\u003cbr\u003e* Trash Classification: The ArmPi Ultra kit includes garbage-pattern blocks. By loading the corresponding model, ArmPi Ultra can quickly identify different types of trash and place them into the correct sorting area.\u003cbr\u003e* Object Sorting: By training models on daily objects and generating corresponding models, and with the support of the depth camera to obtain depth information, ArmPi Ultra can rapidly recognize and pick up the specified items.\u003cbr\u003e\u003cbr\u003eMore Information About YOLO V8:\u003cbr\u003e* Item: Yolo V8\u003cbr\u003e* Backbone: stem 3×3 convolution kernel, bottleneck first-layer expansion\u003cbr\u003e* Accuracy (All Series): 37.3–53.9%\u003cbr\u003e* Inference Speed (T4, TensorRT): 1.92ms\u003cbr\u003e* Supported Tasks: Object detection, instance segmentation, image classification, pose estimation, oriented bounding box (OBB)\u003cbr\u003e* Recommended Application Scenarios: Multi-task fusion, high-precision scenarios, visualization platforms, robotic vision\u003cbr\u003e\u003cbr\u003eMediaPipe Development for Enhanced AI Interaction\u003cbr\u003e* With the MediaPipe development framework, ArmPi Ultra can perform gesture recognition, motion-sensing control, face tracking, and more.\u003cbr\u003e* Gesture Recognition: ArmPi Ultra can recognize different hand gestures, enabling fun and interactive gesture-based control.\u003cbr\u003e* Motion-Sensing Control for Remote Grabbing: By recognizing gestures and determining their positions, ArmPi Ultra can be controlled through motion-sensing to achieve air grabbing without physical touch.\u003cbr\u003e* Face Tracking: ArmPi Ultra can detect faces and obtain their position coordinates. Using inverse kinematics and PID algorithms, it can perform real-time face tracking.\u003cbr\u003e\u003cbr\u003eGazebo Simulation Control\u003cbr\u003e* ArmPi Ultra is developed on the ROS framework and supports GAZEBO simulation. It can simulate robotic-arm control and algorithm validation in a virtual environment, reducing environmental requirements and improving testing efficiency.\u003cbr\u003e\u003cbr\u003eMoveIt Kinematics\u003cbr\u003e* ArmPi Ultra supports MoveIt simulation, enabling robotic-arm control and algorithm verification in a virtual environment, reducing experimental requirements and improving project development efficiency.\u003cbr\u003e* URDF Kinematic Simulation Model: A true-to-scale model of the physical robot. With MoveIt's visual interface, it supports drag-to-move, preset pose control, motion path planning, and obstacle avoidance.\u003cbr\u003e* MoveIt Simulation Control \u0026amp; Motion Planning: The arm can be controlled in real time through the visual simulation interface to perform movements consistent with the physical robot.\u003cbr\u003e* Collision Detection: Obstacles can be added in the simulation environment. As the arm moves, it will automatically avoid collisions.\u003cbr\u003e* Spatial Grabbing: Users can create small custom scenes and drive the robotic arm to perform object picking, placing, and sorting, with full collision detection.\u003cbr\u003e\u003cbr\u003eROS2 Operating System\u003cbr\u003e* Global Mainstream Robotics Communication Framework: ROS2 is the upgraded version of the open-source Robot Operating System ROS1. It inherits all ROS1 functions while supporting more operating systems and compilation environments. ROS2 also enables real-time control and better modular development and testing. Compared with ROS1, it is more powerful and widely applicable.\u003cbr\u003e* Docker Container Technology: Docker is a mainstream container technology that allows developers to package applications into containers and deploy them across different machines quickly and efficiently. Deployment Flow: ROS2 Humble → Docker Engine → Ubuntu 22.04 or other Linux systems → Raspberry Pi 5 motherboard\u003cbr\u003e\u003cbr\u003eCross-Platform Remote Control\u003cbr\u003e* APP Control: App for iOS and Android\u003cbr\u003e* PC Control Software and Wireless Controller Support\u003cbr\u003e\u003cbr\u003e\u003cb\u003eArmPi Ultra Function List:\u003c\/b\u003e\u003cbr\u003e1. Camera\u003cbr\u003e* Target Tracking\u003cbr\u003e* Color Block Tracking\u003cbr\u003e* Color Recognition\u003cbr\u003e* Gesture Interaction\u003cbr\u003e* Fingertip Trajectory Recognition\u003cbr\u003e* MediaPipe Development\u003cbr\u003e* Point Cloud Data Visualization (only for configurations with 3D depth camera)\u003cbr\u003e* Distance Measurement (only for configurations with 3D depth camera)\u003cbr\u003e* 3D Spatial Grasping (only for configurations with 3D depth camera)\u003cbr\u003e\u003cbr\u003e2. ROS Controller (optional Raspberry Pi 5 mainboard)\u003cbr\u003e* Gamepad Control\u003cbr\u003e* VNC Remote Control\u003cbr\u003e* SSH Remote Control\u003cbr\u003e* URDF Model Visualization\u003cbr\u003e* YOLO V8 Model Training\u003cbr\u003e* YOLO V8 Object Detection\u003cbr\u003e* Set AP and STA Modes\u003cbr\u003e* APP Wireless Video Transmission and Control\u003cbr\u003e* Mecanum Wheel Chassis Speed Control\u003cbr\u003e\u003cbr\u003e3. STM32 Controller\u003cbr\u003e* LED Control\u003cbr\u003e* Button Control\u003cbr\u003e* Buzzer Control\u003cbr\u003e* PID Closed-Loop Control\u003cbr\u003e* Gamepad Control\u003cbr\u003e* Robot Serial Communication\u003cbr\u003e* Control Intelligent Bus Servos\u003cbr\u003e* Supports Raspberry Pi 5 Power Protocol\u003cbr\u003e\u003cbr\u003e4. Docker\u003cbr\u003e* Introduction and Installation\u003cbr\u003e* Container Usage\u003cbr\u003e* Docker Repository\u003cbr\u003e* Daily Container Management\u003cbr\u003e* Install ROS in Docker\u003cbr\u003e* Hardware Interaction and Data Processing\u003cbr\u003e* In-depth Understanding and Publishing of Images\u003cbr\u003e\u003cbr\u003e5. AI Voice Interaction Box (advanced version only)\u003cbr\u003e* Voice Playback\u003cbr\u003e* Voice Recognition\u003cbr\u003e* Voice Wake-up\u003cbr\u003e* Custom Wake Words\u003cbr\u003e* Custom Command Words\u003cbr\u003e* Voice-Controlled Color Sorting\u003cbr\u003e* Voice-Controlled Color Tracking\u003cbr\u003e* AI Large Model Voice Interaction\u003cbr\u003e\u003cbr\u003e6. AI Large Model Exclusive Features (advanced version only)\u003cbr\u003e* Deployment and Invocation of Large Language Models\u003cbr\u003e* Semantic Understanding and Emotion Perception\u003cbr\u003e* Deployment and Invocation of Vision Models\u003cbr\u003e* Vision Model Object Recognition\u003cbr\u003e* Vision Model Scene Understanding\u003cbr\u003e* Vision Model Text Recognition\u003cbr\u003e* Overview and Deployment of Multi-Modal Large Models\u003cbr\u003e* Multi-Modal Model Voice Control\u003cbr\u003e* Multi-Modal Color Block Sorting\u003cbr\u003e* Multi-Modal Trash Sorting\u003cbr\u003e* Multi-Modal 3D Object Sorting\u003cbr\u003e* Embodied Intelligence Real-Time Monitoring\u003cbr\u003e* Embodied Intelligence Visual Tracking\u003cbr\u003e* Embodied Intelligence Intelligent Handling\u003cbr\u003e* Embodied Intelligence Smart Assistant\u003cbr\u003e\u003cbr\u003e\u003cb\u003eHigh-Performance Hardware:\u003c\/b\u003e\u003cbr\u003e* Components include an alloy robotic gripper, monocular camera\/ROS depth camera, optional WonderEcho Pro AI voice interaction box, optional Raspberry Pi control system (with heatsink), intelligent bus servos, hard aluminum alloy structure, aluminum alloy chassis, and STM32 controller.\u003cbr\u003e\u003cbr\u003e1. ROS Depth Camera\u003cbr\u003e* The depth camera adopts advanced structured-light 3D imaging technology, integrating a high-performance ISP image signal processor and an ASC depth-computing chip. It requires no additional computing resources while enabling excellent perception capability for robots, making it an ideal choice for 3D vision development projects. It features low power consumption less than 1.6W.\u003cbr\u003e\u003cbr\u003eParameters of the Depth Camera:\u003cbr\u003eModule Specifications: \u003cbr\u003e* Overall Size: 76.5mm × 20.7mm × 21.8mm\u003cbr\u003e* Baseline: 40mm\u003cbr\u003e* Interface: USB 2.0 Wafer connector\u003cbr\u003e* Depth Accuracy: 8mm at 1m\u003cbr\u003e* Depth Precision: 3mm at 0.5m, 7mm at 1m\u003cbr\u003e* Working Distance: 15–300cm\u003cbr\u003e* Operating Temperature: -10°C to 55°C\u003cbr\u003e* Operating Humidity: 0%–95%, non-condensing\u003cbr\u003e* Operating Illumination: 3–80,000Lux\u003cbr\u003e* Power Supply: 5V ±10%, 1.5A\u003cbr\u003e* Device Power Consumption: Average \u0026lt;1.6W\u003cbr\u003e* Safety: Class 1 laser safety\u003cbr\u003e\u003cbr\u003eImaging Performance: \u003cbr\u003e* Depth Data Format: Raw 16-bit\u003cbr\u003e* Depth Resolution \/ Frame Rate: 640 × 400 \/ 12 fps \/ 74° × 51°\u003cbr\u003e* Color Data Format: NV12\u003cbr\u003e* Color Resolution \/ Frame Rate: 640 × 400 \/ 12 fps \/ 74° × 51°\u003cbr\u003e* Infrared Data Format: Raw 8-bit\u003cbr\u003e* Infrared Resolution \/ Frame Rate: 640 × 400 \/ 12 fps \/ 74° × 51\u003cbr\u003e\u003cbr\u003eFirmware Capabilities: \u003cbr\u003e* Firmware Upgrade: Supports USB online upgrade\u003cbr\u003e* Hot Start Delay: \u0026lt;300ms\u003cbr\u003e\u003cbr\u003eSystem Compatibility: \u003cbr\u003e* System Compatibility: For Linux ARMv8, ROS, Windows\u003cbr\u003e\u003cbr\u003e2. WonderEcho Pro AI Voice Interaction Box (Optional)\u003cbr\u003e* ArmPi Ultra is equipped with the WonderEcho Pro voice interaction module, which features a built-in neural network processor and supports custom voice command libraries. It provides the robotic arm with excellent AI voice interaction capability. Combined with vision and sensors, it enables the creation of more advanced and dynamic human–machine interaction projects.\u003cbr\u003e* Product Size: 72mm × 40mm × 23mm\u003cbr\u003e* Power Supply Voltage: DC 5V\u003cbr\u003e* Product Weight: 37g\u003cbr\u003e* Flash Capacity: 2MB\u003cbr\u003e* Recognition Mode: Command wake-up\u003cbr\u003e* Communication Method: Type-C\u003cbr\u003e* Recognition Range: Up to 5m in quiet environments; 1m in noisy environments\u003cbr\u003e* Recognition Requirements: Recognizes fixed vocabulary entries, supporting up to 255 phrases or short sentences, with customizable recognition commands\u003cbr\u003e\u003cbr\u003e3. Full-Perception Self-Feedback Bus Servo\u003cbr\u003e* ArmPi Ultra uses six intelligent bus servos, supporting multi-parameter feedback such as voltage and position to monitor servo status in real time. The servos offer high precision and strong torque, enabling the robotic arm to grasp weights over 500g. Servos are connected in a serial bus structure, ensuring smooth and coordinated multi-joint movements. The design balances engineering aesthetics with dynamic performance.\u003cbr\u003e\u003cbr\u003e\u003cb\u003eProduct Specifications:\u003c\/b\u003e\u003cbr\u003e* Product Name: ArmPi Ultra\u003cbr\u003e* Product Size: 288mm × 177mm × 450mm\u003cbr\u003e* Body Material: Full aluminum alloy body with sandblasted anodized finish\u003cbr\u003e* Controller: STM32F103RBT6\u003cbr\u003e* Camera: USB monocular camera or depth camera\u003cbr\u003e* Power Supply: 7.4V 6A DC\u003cbr\u003e* Operating System: RaspberryOS + Ubuntu 22.04 LTS (Docker)\u003cbr\u003e* Software: PC remote control software + APP for iOS\/Android\u003cbr\u003e* Communication Ways: USB 2.0 ×2; USB 3.0 ×2; Serial port\u003cbr\u003e* Control Methods: ROS Topic\/Node Control; Python Program Control; APP Control; PC Software Control\u003cbr\u003e* Programming Tools: Python, C, C++, Javascript\u003cbr\u003e* Servo Models: HTS-16L; HTS-25L; HX-06L; LX-15D; LX-225 Intelligent Bus Servo\u003cbr\u003e* Supporting Materials: Development manual, video tutorials, ROS source code, STM32 source code, system image, supporting software\u003cbr\u003e\u003cbr\u003e\u003cb\u003ePacking List (Standard Version with Raspberry Pi 5 \u0026amp; SD Card):\u003c\/b\u003e\u003cbr\u003e* 1 × ArmPi Ultra\u003cbr\u003e* 1 × Depth Camera + Data Cable\u003cbr\u003e* 1 × Depth Camera Bracket\u003cbr\u003e* 1 × Raspberry Pi 5 + Power Cable\u003cbr\u003e* 1 × Cooling Fan\u003cbr\u003e* 1 × Raspberry Pi Mini Adapter Board\u003cbr\u003e* 1 × Top Cover + Rear Cover for Main Controller\u003cbr\u003e* 4 × Suction Cups\u003cbr\u003e* 1 × 7.5V 6A Adapter (DC 5.5 × 2.5mm Male)\u003cbr\u003e* 1 × Power Cable (DC 5.5 × 2.1mm Female, 200mm)\u003cbr\u003e* 1 × Wireless Controller\u003cbr\u003e* 1 × 64G Memory Card\u003cbr\u003e* 1 × Card Reader\u003cbr\u003e* 1 × Type-C Data Cable (100mm)\u003cbr\u003e* 1 × Map (600 × 430mm)\u003cbr\u003e* 3 × Tags (30 × 30mm)\u003cbr\u003e* 1 × Set of Trash Cards (40 × 40mm)\u003cbr\u003e* 3 × Colored Blocks (30 × 30mm)\u003cbr\u003e* 4 × Wooden Blocks (40 × 40mm)\u003cbr\u003e* 1 × Rectangular Block (30 × 45mm)\u003cbr\u003e* 1 × Cylinder (30 × 45mm)\u003cbr\u003e* 1 × Roll of Double-Sided Tape (5mm × 8m)\u003cbr\u003e* 1 × Screwdriver\u003cbr\u003e* 1 × Screw Accessory Pack\u003cbr\u003e* 1 × User Manual\u003cbr\u003e\u003cbr\u003e\u003cb\u003ePackaging Details:\u003c\/b\u003e\u003cbr\u003e* Weight: 4.8kg\u003cbr\u003e\u003cb\u003e\u003cspan style=\"color: #ff0000;\"\u003e\u003cbr\u003eNote:\u003cbr\u003e* The robot arm has been assembled.\u003c\/span\u003e\u003c\/b\u003e\u003cbr\u003e\u003cbr\u003e\u003cb\u003eDownload Tutorials \u0026amp; Software:\u003c\/b\u003e\u003cbr\u003e\u003ca href=\"https:\/\/drive.google.com\/drive\/folders\/1UCUnH4N7RXv_0zCiUpTw__6oypUB5E2H?usp=sharing\" target=\"_blank\"\u003ehttps:\/\/drive.google.com\/drive\/folders\/1UCUnH4N7RXv_0zCiUpTw__6oypUB5E2H?usp=sharing\u003c\/a\u003e\u003cbr\u003e\u003cbr\u003e\u003cimg style=\"width: 1008px;\" src=\"https:\/\/www.comseek.com\/image\/catalog\/99844\/99844-01.jpg\"\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cimg style=\"width: 1008px;\" src=\"https:\/\/www.comseek.com\/image\/catalog\/99844\/99844-02.jpg\"\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cimg style=\"width: 1008px;\" src=\"https:\/\/www.comseek.com\/image\/catalog\/99844\/99844-03.jpg\"\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cimg style=\"width: 1008px;\" src=\"https:\/\/www.comseek.com\/image\/catalog\/99844\/99844-04.jpg\"\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cimg style=\"width: 1008px;\" src=\"https:\/\/www.comseek.com\/image\/catalog\/99844\/99844-05.jpg\"\u003e\u003cbr\u003e\u003c\/p\u003e","brand":"Hiwonder","offers":[{"title":"Default Title","offer_id":41028775673905,"sku":"99851","price":711.9,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0578\/4790\/6353\/files\/1763429646-1.jpg?v=1780336850","url":"https:\/\/www.thanksbuyer.com\/ko\/products\/hiwonder-armpi-ultra-robot-arm-standard-version-with-depth-camera-raspberry-pi-5-8gb-amp-sd-card","provider":"Thanksbuyer Online,SDR Radio, Testers and Measurement devices, Astronomy products,  Audio Amplifer and Decoders","version":"1.0","type":"link"}