Our Products

Discover our range of quadruped robots designed for various applications

Panda Agent Collaborative Arm Vision Kit

Panda Agent Collaborative Arm Vision Kit

6-axis collaborative arm / RGBD vision

{"intro":"Centered on an ultra-lightweight 6-axis collaborative robotic arm, integrated with an AI electric gripper, RGBD depth camera, and 40T AI edge computing box, along with motion control, vision guidance, and AI algorithm software, this forms an out-of-the-box robotic visual grasping platform. The whole unit weighs about 8kg, flexible for installation on mobile platforms or fixed stands.","headers":{"intro":"Product Overview","features":"Key Features","specs":"Technical Specifications","scenarios":"Typical Applications"},"features":[{"label":"6-axis Collaborative Arm","value":"Payload 5kg, reach 660mm, precision ±0.05mm, supports drag teaching"},{"label":"AI Electric Gripper","value":"Opening 0~100mm, adjustable gripping force, RS485/Modbus RTU communication"},{"label":"RGBD Depth Camera","value":"Depth 640×480@30fps, RGB 1280×720, range 0.6~8m"},{"label":"40T Computing Box","value":"40 TOPS computing power, 32GB memory, supports TensorRT/CUDA, local inference with zero cloud dependency"},{"label":"Vision Guidance","value":"Hand-eye calibration translation error ≤0.82mm, supports 6D pose estimation and grasp planning"},{"label":"Motion Control","value":"4ms communication cycle, Ethernet TCP, with Python SDK v4.0"},{"label":"AI Algorithm Package","value":"Object detection/classification, 6D pose estimation, grasp point planning"},{"label":"Development Support","value":"Python SDK, sample code, technical documentation, 1-year manufacturer support"}],"specs":[{"label":"Degrees of Freedom","value":"6 axes (J1 base rotation ±360°, J2~J5 all joints with brakes)"},{"label":"Payload","value":"5kg (rated)"},{"label":"Reach","value":"Horizontal approx. 660mm, vertical 780mm"},{"label":"Repeatability","value":"±0.05mm"},{"label":"Control Interface","value":"Ethernet TCP, communication cycle 4ms"},{"label":"Computing Power","value":"≥40 TOPS, 12-core CPU, 1792-core GPU, runs Ubuntu 20.04"}],"scenarios":["Quadruped robot inspection operations","Valve/switch operation","Equipment status collection","Abnormal handling","Material transfer"]}

6-axis collaborative armRGBD vision40T edge computingGrasp planning
Learn More

Panda Agent Speech-to-Text System

Speech Recognition / Offline AI

{"intro":"Based on the FasterWhisper Small model, it provides fully offline Chinese speech-to-text capability. Built-in VAD endpoint detection for automatic sentence segmentation, no internet required, supports CPU/CUDA inference, suitable for quadruped robot command parsing, industrial equipment voice control, intelligent security voice recording, and other scenarios.","headers":{"intro":"Product Overview","features":"Key Features","specs":"Technical Specifications","scenarios":"Typical Applications"},"features":[{"label":"Speech-to-Text","value":"Real-time conversion of microphone input to Chinese text, supporting continuous recognition"},{"label":"VAD Segmentation","value":"Silence threshold 800ms automatically detects speech boundaries, no manual trigger needed"},{"label":"Fully Offline","value":"Local inference, zero network dependency, data stays on edge device"},{"label":"int8 Quantization","value":"Reduces computation, improves CPU utilization efficiency"},{"label":"Multi-core Parallelism","value":"Automatically utilizes multiple logical CPU cores"},{"label":"USB Microphone","value":"Supports standard USB audio devices, 16kHz mono PCM capture"},{"label":"Result Output","value":"Text results output to application layer in real time, supports JSONL logging"},{"label":"CUDA Acceleration","value":"Switchable CUDA inference (Jetson Orin), significantly reduced latency"}],"specs":[{"label":"STT Model","value":"FasterWhisper Small (CTranslate2 int8 quantized)"},{"label":"Language Support","value":"Chinese (Mandarin), extensible to multilingual"},{"label":"Sampling Parameters","value":"16000Hz mono int16 PCM"},{"label":"VAD Threshold","value":"Silence 800ms triggers segmentation"},{"label":"Runtime Memory","value":"Resident ~405MiB, peak ~544MiB"},{"label":"Runtime Environment","value":"Python 3.8+, faster-whisper, Ubuntu 20.04"}],"scenarios":["Quadruped Robot Voice Commands","Industrial Equipment Voice Control","Security Voice Recording","Smart Q&A Frontend","Voice Memos"]}

Speech RecognitionOffline AIEdge ComputingVAD Segmentation
Learn More
Panda Agent UWB Intelligent Human Following System

Panda Agent UWB Intelligent Human Following System

UWB following / AOA positioning

{"intro":"Based on Ultra-Wideband (UWB) AOA angle-of-arrival positioning technology, it enables real-time autonomous following of a specific person by the robot. No camera required, unaffected by lighting or obstructions, stable operation in complex indoor and outdoor environments. Compatible with mainstream quadruped robot platforms such as Unitree, Beijing Zhishen, Tianjin Galio, Shanghai Zhiyuan, and Deep Robotics.","headers":{"intro":"Product Overview","features":"Key Features","specs":"Technical Specifications","scenarios":"Typical Applications"},"features":[{"label":"Automatic Following","value":"Real-time target perception, automatically adjusts speed/direction to maintain set distance"},{"label":"Adjustable Distance","value":"1~20m dynamic adjustment, default 2m, one-click modification via Web interface"},{"label":"Direction Modes","value":"Four following directions: front/back/left/right, one-click switch via Web or remote control"},{"label":"Proximity Protection","value":"Auto stop/reverse when target is too close to prevent collision (stops at ≤0.6m)"},{"label":"Physical Remote Control","value":"Metal power switch + four buttons (red/yellow1/yellow2/green) for mode switching and manual control"},{"label":"Disconnection Protection","value":"Auto stop after 1s UWB timeout, auto resume on reconnection"},{"label":"Web Control","value":"One-click start/stop via browser, real-time display of distance, azimuth, and status"},{"label":"Three Speed Levels","value":"Low 0.4m/s, Normal 0.8m/s, High 1.2m/s"}],"specs":[{"label":"Positioning Method","value":"UWB AOA (simultaneous distance and azimuth acquisition)"},{"label":"Ranging Range","value":"0.1~30m, recommended following range 1~20m"},{"label":"Data Frequency","value":"20Hz control loop"},{"label":"Interface","value":"USB Serial 921600 baud rate + UDP motion commands"},{"label":"Compatible Platforms","value":"Unitree Go/B/H, Beijing Zhishen, Tianjin Galio, Shanghai Zhiyuan, Deep Robotics"}],"scenarios":["Security patrol following","Material handling assistance","Exhibition tour guidance","Warehouse logistics accompaniment"]}

UWB followingAOA positioningAutonomous following
Learn More

Panda Agent Voiceprint Recognition Algorithm System

Voiceprint Recognition / Offline AI

{"intro":"Based on the WeSpeaker ResNet34 deep neural network model, it provides fully offline speaker enrollment and recognition capabilities. No internet connection required, supports local inference on edge devices, with recognition latency as low as 84ms (for 5-second speech). Suitable for scenarios with strict privacy and real-time requirements such as quadruped robots, industrial equipment, and smart access control.","headers":{"intro":"Product Overview","features":"Key Features","specs":"Technical Specifications","scenarios":"Typical Applications"},"features":[{"label":"Speaker Enrollment","value":"Capture ≥3s speech, extract 256-dimensional voiceprint feature vector and store in database"},{"label":"Speaker Recognition","value":"Compare with voiceprint database, output best match and similarity score"},{"label":"Multi-user Voiceprint Database","value":"Supports simultaneous enrollment of multiple users, verified for ≥4 concurrent recognition"},{"label":"Fully Offline","value":"Local ONNX inference, zero network dependency, data never leaves edge device"},{"label":"Edge Deployment","value":"Supports Jetson Orin CPU/CUDA/TensorRT"},{"label":"Real-time Response","value":"RTF approx. 0.045, processing speed about 22 times real-time"},{"label":"USB Microphone","value":"Supports 16kHz mono audio capture, plug-and-play without drivers"},{"label":"Result Announcement","value":"Recognition results output in real-time via USB speaker or UART"}],"specs":[{"label":"Base Model","value":"WeSpeaker ResNet34 CnCeleb (ONNX float32, 25.3MB)"},{"label":"Feature Dimension","value":"256-dimensional Speaker Embedding"},{"label":"Inference Latency","value":"5s audio CPU single-thread 227ms / 2-thread 134ms / 4-thread 91ms"},{"label":"RTF","value":"Approx. 0.045"},{"label":"Runtime Environment","value":"Python 3.8+, sherpa-onnx, ONNX Runtime, Ubuntu 20.04"}],"scenarios":["Quadruped Robot Identity Authentication","Industrial Equipment Voiceprint Access Control","Smart Security","Voice Command Authorization","Attendance Verification"]}

Voiceprint RecognitionOffline AIEdge Deployment
Learn More

产品展示

Panda Agent Collaborative Arm Vision Kit 1

Panda Agent Collaborative Arm Vision Kit

Panda Agent Collaborative Arm Vision Kit - 40T Edge AI Box 1

Panda Agent Collaborative Arm Vision Kit - 40T Edge AI Box