Skip to content

🤖 PND Adam Humanoid Robot

humanoid Adam is a full-body robotic system that integrates cutting-edge hardware architecture with an intelligent algorithm platform, offering exceptional motion flexibility, environmental perception, and high-precision dynamic control. Its bionic structure design and human-like motion abilities enable it to perform highly complex tasks in unstructured environments. It is widely applicable across various cutting-edge fields including scientific research experiments, industrial automation testing, medical assistance , and service robotics.


💡 Key Features

🦾 Mechanical System Architecture

  • Actuation System: Integrated with up to 41 quasi-direct drive (Quasi-Direct Drive) flexible force-controlled joints, enabling high dynamic precision control and multi-degree-of-freedom dexterous movement.
  • Modular Architecture: Features a fully modular actuation system, optimizing production and maintenance efficiency, and meeting large-scale manufacturing needs.
  • Bionic Hip Joint Design: An innovative bionic hip structure that significantly enhances the robot's human-like motion and flexibility in complex operational tasks.
  • Lower Limb Joint System: Equipped with 4 high-torque density, multi-stage reduction ratio, high-response sensitivity quasi-direct drive force-controlled joints, providing excellent gait performance and dynamic stability.

🧠 Control System & Communication Architecture

  • PND-Network: Proprietary high-real-time communication protocol stack, supporting the integration of multimodal sensors such as LiDAR and visual cameras, enhancing environmental perception and closed-loop control.
  • Central Control Unit: Powered by the Intel NUC12WSKi7 processor and a self-developed real-time control unit (RCU), coordinating the robot's joint drives, energy management, communication scheduling, and system monitoring.
  • AI Computing Platform: The Adam Standard/SP integrates the NVIDIA Jetson Orin NX 16GB embedded module, specialized in high-performance AI inference, visual computing, and real-time decision-making tasks.
  • Perception System: Equipped with the Intel RealSense D455 depth vision sensor, achieving high-precision 3D environmental modeling and real-time spatial positioning.

🌀 Dynamic Balance & Motion Planning

  • WBC + MPC Algorithm Architecture: A fusion strategy of Whole-Body Control (WBC) and Model Predictive Control (MPC), ensuring the robot's balance robustness and trajectory tracking accuracy in dynamic disturbance environments.
  • Reinforcement Learning & Imitation Learning: Utilizes large-scale simulation environments and deep neural network training, combining reinforcement learning and imitation learning strategies to iteratively optimize motion strategies and control performance.

🔋 Energy & Endurance Management

  • Intelligent Battery Management System (BMS): Integrated high-precision power monitoring and dynamic power distribution mechanisms to support continuous high-intensity task execution.
  • System Energy Efficiency Optimization: Low-power hardware design and dynamic power consumption management strategies significantly enhance the overall system's endurance.

📡 Communication & System Expansion

  • Multimodal Communication Capabilities: Compatible with 5G, Wi-Fi 6, Bluetooth 5.0 protocols for high-bandwidth, low-latency remote control and real-time debugging.
  • Sensor Expansion Interface: PND-Network features excellent interface expandability, supporting multi-source sensor fusion and system functionality customization, empowering complex scene perception and decision-making capabilities.

🌐 Application Fields

  • Scientific Research & Laboratory Scenarios: Suitable for high-precision automated experiments, dynamic system testing, and data acquisition analysis.
  • Industrial Automation Production Lines: Achieves high-reliability operations and process optimization in precision assembly, online inspection, and flexible manufacturing.
  • Medical Assistance & Rehabilitation: Provides surgical collaboration, rehabilitation training support, and patient monitoring services, adapting to diverse healthcare scenarios.
  • Service & Collaborative Robotics: Applicable in logistics sorting, reception guidance, human-robot collaboration, and other intelligent service applications, with high adaptability to various environments.

📊 Technical Specifications

Robot Model Adam Lite Adam Standard Adam SP Adam Pro
Body Size 1.67m 1.67m 1.67m 1.67m
Total Weight 60kg 61kg 62kg 63kg
Total Degrees of Freedom 25 degrees of freedom 29 degrees of freedom 41 degrees of freedom 43 degrees of freedom
Leg Degrees of Freedom Hip x3 + Knee x1 + Ankle x2 = 6 Hip x3 + Knee x1 + Ankle x2 = 6 Hip x3 + Knee x1 + Ankle x2 = 6 Hip x3 + Knee x1 + Ankle x2 = 6
Arm Degrees of Freedom Shoulder x3 + Elbow x1 + Forearm x1 = 5 Shoulder x3 + Elbow x1 + Forearm x1 + Wrist x2 = 7 Shoulder x3 + Elbow x1 + Forearm x1 + Wrist x2 + Dexterous Fingers x6 = 13 Shoulder x3 + Elbow x1 + Forearm x1 + Wrist x2 + Dexterous Fingers x6 = 13
Waist Degrees of Freedom 3 degrees of freedom 3 degrees of freedom 3 degrees of freedom 3 degrees of freedom
Joint Torque Limits Knee ~ 340N·m, Hip ~ 340N·m, Ankle ~ 46N·m, Arm ~ 60N·m Knee ~ 340N·m, Hip ~ 340N·m, Ankle ~ 46N·m, Arm ~ 60N·m Knee ~ 340N·m, Hip ~ 340N·m, Ankle ~ 46N·m, Arm ~ 60N·m Knee ~ 340N·m, Hip ~ 340N·m, Ankle ~ 46N·m, Arm ~ 60N·m
Walking Speed Max 1.5m/s Max 1.5m/s Max 1.5m/s Max 1.5m/s
Battery 1188W·h capacity, 46.2V max voltage, 25A max output current 1188W·h capacity, 46.2V max voltage, 25A max output current 1188W·h capacity, 46.2V max voltage, 25A max output current 1188W·h capacity, 46.2V max voltage, 25A max output current
Communication Wi-Fi 6, Bluetooth 5.0, easy remote real-time control and debugging Wi-Fi 6, Bluetooth 5.0, easy remote real-time control and debugging Wi-Fi 6, Bluetooth 5.0, easy remote real-time control and debugging Wi-Fi 6, Bluetooth 5.0, easy remote real-time control and debugging
Control & Perception Computing Robust self-developed WBC (Whole Body Control) + MPC (Model Predictive Control) algorithms for dynamic balance Robust self-developed WBC (Whole Body Control) + MPC (Model Predictive Control) algorithms for dynamic balance Robust self-developed WBC (Whole Body Control) + MPC (Model Predictive Control) algorithms for dynamic balance Robust self-developed WBC (Whole Body Control) + MPC (Model Predictive Control) algorithms for dynamic balance
Motion Control Computer NUC12WSKi7 for motion control NUC12WSKi7 for motion control NUC12WSKi7 for motion control NUC12WSKi7 for motion control
Perception & AI Computer / NVIDIA Jetson Orin NX 16GB embedded computer for AI, vision, and decision making NVIDIA Jetson Orin NX 16GB embedded computer for AI, vision, and decision making NVIDIA Jetson Orin NX 16GB embedded computer for AI, vision, and decision making
Perception Sensors / Intel Realsense D455 depth camera Intel Realsense D455 depth camera ZED MINI
Head / 2 degrees of freedom
Hand Spherical Hand / Dexterous Hand Dexterous Hand

🧍‍♂️ Appearance Overview

🧩 Key Components

Component Name Standard Version Plus Version Pro Version



The head is for display purposes only, for reference.

⚙️ Hardware Architecture

Hardware Architecture


🧠 Control & Compute Modules

🧮 Motion Control Unit (Cerebellum)

Parameter Specification
Model NUC12WSKi7
CPU 12th Gen Intel Core i7-1260P
Cores 12 cores
Threads 16 threads
Max Turbo Frequency 4.7GHz
RAM 16GB
RAM Type DDR4 3200M Hz
Cache 18MB
Storage 128GB
GPU Intel® Iris® Xe Graphics
Max Dynamic GPU Frequency 1.40 GHz
Intel® Deep Learning Boost Yes
Intel® Adaptix™ Technology Yes
Intel® Hyper-Threading Technology Yes
Instruction Set 64bit

🧬 Perception & Decision Unit (Brain)

Parameter Specification
Model NVIDIA Jetson Orin NX 16GB
AI Performance 100 TOPS
GPU 32 Tensor Cores, 1024-core NVIDIA Ampere architecture GPU
CPU 8-core Arm CortexA78AE v8.2 64-bit CPU 2MB L2 + 4MB L3
Memory 16GB 128-bit LPDDR5, 102.4GB/s
Storage 128GB M.2 SSD
Video Encoding 1x 4K60 (H.265), 3x 4K30 (H.265), 6x 1080p60 (H.265), 12x 1080p30 (H.265)
Video Decoding 1x 8K30 (H.265), 2x 4K60 (H.265), 4x 4K30 (H.265), 9x 1080p60 (H.265), 18x 1080p30 (H.265)
Camera 2x MIPI CSI-2 D-PHY channels
USB 4x USB 3.2, 1x USB Type-C
Display Interface 1x DisplayPort
Networking Gigabit Ethernet
Other I/O 40-pin header (UART, SPI, I2S, I2C, PWM, GPIO), 4-pin fan header, DC power socket
Size 103x90x34 (mm)

🔩 Joint Parameters

🤖 Adam Lite

Joint Name IP Address Actuator Model Limit (rad) Notes
shoulderPitch_Left/Right 10.10.10.10 / 10.10.10.30 PND-50-14A-50-S +2.0420~-3.6138 Positive for backward, negative for forward
shoulderRoll_Left/Right 10.10.10.11 / 10.10.10.31 PND-50-14A-50-S Left: +3.1416~-0.6283
Right: +0.6283~-3.1416
Left: Positive outward, negative inward
Right: Positive inward, negative outward
shoulderYaw_Left/Right 10.10.10.12 / 10.10.10.32 PND-30-14A-50-S ±2.5831
elbow_Left/Right 10.10.10.13 / 10.10.10.33 PND-30-14A-50-S +0.2094~-2.4958 Positive for backward, negative for forward
wristYaw_Left/Right 10.10.10.14 / 10.10.10.34 PND-20-14A-50-S ±2.6704
waistRoll 10.10.10.90 PND-60-17-50-S ±0.2793
waistPitch 10.10.10.91 PND-60-17-50-S +1.3614~-0.8378 Positive for forward, negative for backward
waistYaw 10.10.10.92 PND-60-17-50-S ±0.8290
hipPitch_Left/Right 10.10.10.70 / 10.10.10.50 PND-130-92-7-P ±2.2276
hipRoll_Left/Right 10.10.10.71 / 10.10.10.51 PND-80-20-30-S +1.6581~-0.7854 Positive outward, negative inward
hipYaw_Left/Right 10.10.10.72 / 10.10.10.52 PND-60-17-50-S ±0.8290
kneePitch_Left/Right 10.10.10.73 / 10.10.10.53 PND-130-92-7-P +2.4435~0.0000
anklePitch_Left/Right 10.10.10.74 / 10.10.10.54 PND-50-52-30-P +0.4276~-1.0472 Two joints linked, toe-down is negative, reverse is positive
ankleRoll_Left/Right 10.10.10.75 / 10.10.10.55 PND-50-52-30-P ±0.4887 Two joints linked

Lite

🤖 Adam Standard

Joint Name IP Address Actuator Model Limit (rad) Notes
shoulderPitch_Left/Right 10.10.10.10 / 10.10.10.30 PND-50-14A-50-S +2.0420~-3.6138 Positive for backward, negative for forward
shoulderRoll_Left/Right 10.10.10.11 / 10.10.10.31 PND-50-14A-50-S Left: +2.7925~-0.6283
Right: +0.6283~-2.7925
Left: Positive outward, negative inward
Right: Positive inward, negative outward
shoulderYaw_Left/Right 10.10.10.12 / 10.10.10.32 PND-30-14A-50-S ±2.5831
elbow_Left/Right 10.10.10.13 / 10.10.10.33 PND-30-14A-50-S +0.2094~-2.4958 Positive for backward, negative for forward
wristYaw_Left/Right 10.10.10.14 / 10.10.10.34 PND-20-14A-50-S ±2.6704
wristPitch_Left/Right 10.10.10.15 / 10.10.10.35 PND-20-08-50-S ±0.9599 Two joints linked, spherical joint range of motion
wristRoll_Left/Right 10.10.10.16 / 10.10.10.36 PND-20-08-50-S ±0.9599 Two joints linked, spherical joint range of motion
waistRoll 10.10.10.90 PND-60-17-50-S ±0.2793
waistPitch 10.10.10.91 PND-60-17-50-S +1.3614~-0.8378 Positive for forward, negative for backward
waistYaw 10.10.10.92 PND-60-17-50-S ±0.8290
hipPitch_Left/Right 10.10.10.70 / 10.10.10.50 PND-130-92-7-P ±2.2276
hipRoll_Left/Right 10.10.10.71 / 10.10.10.51 PND-80-20-30-S +1.6581~-0.7854 Positive outward, negative inward
hipYaw_Left/Right 10.10.10.72 / 10.10.10.52 PND-60-17-50-S ±0.8290
kneePitch_Left/Right 10.10.10.73 / 10.10.10.53 PND-130-92-7-P +2.4435~0.0000
anklePitch_Left/Right 10.10.10.74 / 10.10.10.54 PND-50-52-30-P +0.4276~-1.0472 Two joints linked, toe-down is negative, reverse is positive
ankleRoll_Left/Right 10.10.10.75 / 10.10.10.55 PND-50-52-30-P ±0.4887 Two joints linked

Standard

🤖 Adam SP

Joint Name IP Address Actuator Model Limit (rad) Notes
shoulderPitch_Left/Right 10.10.10.10 / 10.10.10.30 PND-50-14A-50-S +2.0420~-3.6138 Positive for backward, negative for forward
shoulderRoll_Left/Right 10.10.10.11 / 10.10.10.31 PND-50-14A-50-S Left: +2.7925~-0.6283
Right: +0.6283~-2.7925
Left: Positive outward, negative inward
Right: Positive inward, negative outward
shoulderYaw_Left/Right 10.10.10.12 / 10.10.10.32 PND-30-14A-50-S ±2.5831
elbow_Left/Right 10.10.10.13 / 10.10.10.33 PND-30-14A-50-S +0.2094~-2.4958 Positive for backward, negative for forward
wristYaw_Left/Right 10.10.10.14 / 10.10.10.34 PND-20-14A-50-S ±2.6704
wristPitch_Left/Right 10.10.10.15 / 10.10.10.35 PND-20-08-50-S ±0.9599 Two joints linked, spherical joint range of motion
wristRoll_Left/Right 10.10.10.16 / 10.10.10.36 PND-20-08-50-S ±0.9599 Two joints linked, spherical joint range of motion
waistRoll 10.10.10.90 PND-60-17-50-S ±0.2793
waistPitch 10.10.10.91 PND-60-17-50-S +1.3614~-0.8378 Positive for forward, negative for backward
waistYaw 10.10.10.92 PND-60-17-50-S ±0.8290
hipPitch_Left/Right 10.10.10.70 / 10.10.10.50 PND-130-92-7-P ±2.2276
hipRoll_Left/Right 10.10.10.71 / 10.10.10.51 PND-80-20-30-S +1.6581~-0.7854 Positive outward, negative inward
hipYaw_Left/Right 10.10.10.72 / 10.10.10.52 PND-60-17-50-S ±0.8290
kneePitch_Left/Right 10.10.10.73 / 10.10.10.53 PND-130-92-7-P +2.4435~0.0000
anklePitch_Left/Right 10.10.10.74 / 10.10.10.54 PND-50-52-30-P +0.4276~-1.0472 Two joints linked, toe-down is negative, reverse is positive
ankleRoll_Left/Right 10.10.10.75 / 10.10.10.55 PND-50-52-30-P ±0.4887 Two joints linked
hand_Left/Right 10.10.10.18 / 10.10.10.38 / / Dexterous hand joint

SP

🤖 Adam Pro

Joint Name IP Address Actuator Model Limit (rad) Notes
shoulderPitch_Left/Right 10.10.10.10 / 10.10.10.30 PND-50-14A-50-S +2.0420~-3.6138 Positive for backward, negative for forward
shoulderRoll_Left/Right 10.10.10.11 / 10.10.10.31 PND-50-14A-50-S Left: +2.7925~-0.6283
Right: +0.6283~-2.7925
Left: Positive outward, negative inward
Right: Positive inward, negative outward
shoulderYaw_Left/Right 10.10.10.12 / 10.10.10.32 PND-30-14A-50-S ±2.5831
elbow_Left/Right 10.10.10.13 / 10.10.10.33 PND-30-14A-50-S +0.2094~-2.4958 Positive for backward, negative for forward
wristYaw_Left/Right 10.10.10.14 / 10.10.10.34 PND-20-14A-50-S ±2.6704
wristPitch_Left/Right 10.10.10.15 / 10.10.10.35 PND-20-08-50-S ±0.9599 Two joints linked, spherical joint range of motion
wristRoll_Left/Right 10.10.10.16 / 10.10.10.36 PND-20-08-50-S ±0.9599 Two joints linked, spherical joint range of motion
waistRoll 10.10.10.90 PND-60-17-50-S ±0.2793
waistPitch 10.10.10.91 PND-60-17-50-S +1.3614~-0.8378 Positive for forward, negative for backward
waistYaw 10.10.10.92 PND-60-17-50-S ±0.8290
hipPitch_Left/Right 10.10.10.70 / 10.10.10.50 PND-130-92-7-P ±2.2276
hipRoll_Left/Right 10.10.10.71 / 10.10.10.51 PND-80-20-30-S +1.6581~-0.7854 Positive outward, negative inward
hipYaw_Left/Right 10.10.10.72 / 10.10.10.52 PND-60-17-50-S ±0.8290
kneePitch_Left/Right 10.10.10.73 / 10.10.10.53 PND-130-92-7-P +2.4435~0.0000
anklePitch_Left/Right 10.10.10.74 / 10.10.10.54 PND-50-52-30-P +0.4276~-1.0472 Two joints linked, toe-down is negative, reverse is positive
ankleRoll_Left/Right 10.10.10.75 / 10.10.10.55 PND-50-52-30-P ±0.4887 Two joints linked
hand_Left/Right 10.10.10.18 / 10.10.10.38 / / Dexterous hand joint
neckYaw 10.10.10.93 PND-60-17-50-S ±1.0472 Neck yaw joint
neckPitch 10.10.10.94 PND-60-17-50-S ±1.0472 Neck pitch joint

Pro

✋ PND Dexterous Hand

Joint Name Limit (rad) Notes
Thumb1_Left/Right 0.0873 ~ 1.5708 Abduction
Thumb2_Left/Right 0.4538 ~ 1.0821 Flexion
Index_Left/Right 0 ~ 1.5533 Flexion
Middle_Left/Right 0 ~ 1.5533 Flexion
Ring_Left/Right 0 ~ 1.5533 Flexion
Pinky_Left/Right 0 ~ 1.5533 Flexion

hands

The finger joints adopt a linear control mapping:

A control value of 0 corresponds to the joint being at the fully open position (i.e., the lower angle limit).

A control value of 1000 corresponds to the joint being at the fully closed position (i.e., the upper angle limit).


Last Updated: 2025-12-03