news

Traditional robotic arms began with repetitive tasks in automotive and electronics manufacturing, driving industrial automation. With sensors and machine vision, they became more adaptable to dynamic environments, paving the way for humanoid robots. Recent advances in AGI research and machine learning have enabled humanoid robots to learn dynamically and adapt in real-time, unlocking applications in services, healthcare, and home assistance.

• What is Artificial General Intelligence ?

Artificial General Intelligence (AGI) is a yet-to-be-realized system viewed as a key goal in AI development. Unlike Narrow AI, such as GPT specialized in NLP, AGI aims for human-like learning, reasoning, and adaptability across diverse contexts. While AGI faces challenges like general reasoning, cross-domain learning, and unknown scenario adaptation, humanoid robots provide a promising platform for exploration. Equipped with AI, these robots mimic human behavior, perceive surroundings, and act autonomously, serving as valuable testbeds for integrating AGI into real-world applications.

• What are the Core Components of Humanoid Robots?

Humanoid robots are primarily composed of four key modules: sensing, control, execution, and power modules. These modules work collaboratively, not only enabling robots to simulate human movements but also gradually achieving autonomous actions. In industrial applications, these robots are already capable of performing assembly and transportation tasks in complex automotive factories, significantly improving production efficiency and reducing human labor.

• Sensing Module

The sensing module is responsible for collecting environmental data surrounding the robot, including images, sound, temperature, and positional information. This data is captured through various sensors, such as vision, tactile, audio, and pressure modules. Since over 80% of human knowledge is acquired through vision, the visual system plays a critical role in humanoid robots.

For example, Tesla Optimus, with its 8 cameras, can construct 3D objects and environmental information. By integrating devices such as LiDAR, depth cameras, and ultrasonic sensors, it achieves multimodal fusion, enabling precise environmental perception through Occupancy Networks, allowing robots to respond swiftly in dynamic scenarios.

To enable these capabilities, data transmission plays a vital role in ensuring that diverse sensor data is transmitted quickly and reliably to the central processing unit. Additionally, to meet the spatial constraints and multi-sensor configuration demands of humanoid robots,

Sensing module connection solutions must exhibit the following three key features: 

1. High-Speed Analog/Digital Signal Transmission:

Regardless of how advanced the AI algorithms of a humanoid robot may be, they rely on high-speed connection solutions to deliver high-resolution analog signals or process data from dozens of sensors.

2. Effective Shielding: 

Signal lines not only need to carry high-speed signals but also share limited space with power system wiring. Effective shielding is crucial to prevent internal interference and to mitigate the robot's impact on the external environment during operation.

3. Lightweight and High-Strength Design: 

Lightweight connection solutions improve the mobility of humanoid robots and extend their battery life. They must also consider the impact of vibrations on signal transmission during robot movement.

• 
• Control Module

The control module serves as the intelligent core of a robot, responsible for decision-making and motion control. It integrates high-performance chips such as CPUs, GPUs, or AI accelerators. Beyond hardware computational capabilities, AI-driven software is even more critical, as it enables the robot to make autonomous decisions, learn independently, and adapt to its environment.

Large language models like GPT interpret unstructured data from sensor modules, such as images and audio, analyze context, and generate commands. Reinforcement learning helps robots adapt to dynamic environments and optimize decisions through trial and error. Other AI algorithms, like path planning and real-time decision-making, translate commands into actions executed by the actuator module. For example, RL can optimize greetings when someone approaches or guide robots to provide comfort based on emotional cues, enabling more natural interactions.

Unlike standard industrial computers, the control module of humanoid robots must address three key characteristics:

1. Customized Compact Design: 

The limited internal space of humanoid robots demands highly customized controllers tailored to specific configurations, considering lightweight structures and interface arrangements

2. Vibration-Resistant Design: 

The vibrations generated by humanoid robots during movement often exceed those in vehicles. The structural integrity of the control module must account for vibration resistance and environmental durability.

3. Thermal Management Design:   

Humanoid robots can only use fanless control modules, which must address the heat accumulation from high-performance chips and algorithmic computations. Efficient fanless thermal management is crucial to prevent overheating and maintain system performance.

• Execution Module

The execution module is responsible for converting the commands from the control module into specific actions, such as walking, lifting, and performing precise operations like rotating. This module comprises various actuators (e.g., servo motors and hydraulic devices), control and drive systems (e.g., servo drives), auxiliary mechanical components (e.g., reducers, casings, and bearings), and position feedback elements (e.g., encoders). Servo drives regulate motor output based on control signals, while reducers (such as harmonic reducers) enhance torque and precision. Encoders continuously monitor movement positions to provide accurate feedback. The collaboration of these high-precision components enables robots to mimic fundamental human motion patterns, delivering flexible and stable movement performance.

The connection systems and mechanical components in the execution module must endure high-frequency dynamic stress. To meet the demands of high loads and frequent movements in robots, in addition to the core drive system.

In the Execution Module, the following issues with mechanical parts and connection systems need to be addressed:

1. Lightweight and High-Strength Connection Systems:

Using specialized materials that reduce weight while maintaining high strength, capable of withstanding stress during movement and repeated multi-angle bending at joint areas.

2. Diverse Connection Solutions:

The cables and connectors in humanoid robots require distinct solutions across different execution modules. For instance, the power transmission needs of the torso differ entirely from the flexibility and elasticity demands of hand joints. Diverse connection solutions are critical, including the selection of cables and the design of connectors, to ensure optimal performance for various requirements.

• Power Module

The power module serves as the core driving force of the robot, providing stable electricity to ensure sufficient energy for all modules performing various tasks. It comprises a Battery Management System, inverters, Power Distribution Units , and a cooling system, dynamically distributing power according to task loads. This ensures the stable operation of critical components and prevents operational interruptions. Depending on application requirements, power replenishment methods include:

Battery Replacement:  Ideal for high-intensity, long-duration tasks, allowing for quick battery swaps to minimize downtime.

Charging Stations: Suitable for fixed-range applications (e.g., automated delivery or cleaning), enabling robots to autonomously navigate to charging stations during task intervals.

Wired Charging: Designed for overnight or downtime use, providing stable and consistent power supply.

During the charging process, connectors play a critical role in ensuring stable and efficient power transmission to the power module while preventing external interference or overheating that could impact charging efficiency. To ensure the stability and safety of the charging process,

Power module connectors must exhibit the following four key characteristics: 

1. High Current Capacity:

Capable of stable operation under high currents, preventing overheating and performance degradation.

2. High-Temperature Resistance and Fire-Retardant Design:

Ensures safety during prolonged operation and reduces the risk of short circuits.

3. Low Impedance and Thermal Management Design:

Minimizes energy loss, reduces heat generation, and enhances power transmission efficiency.

4. High Durability and Vibration Resistance:

Maintains stable connections under repeated plug-and-unplug operations and vibration conditions. Additionally, connectors must feature waterproof and dustproof capabilities to extend their service life.

Conclusion

Humanoid robots lag behind Tesla's autonomous driving in commercial maturity but are advancing steadily with great potential. Elon Musk predicts 10 billion humanoid robots by 2040, outnumbering humans, as technology advances and costs decline, paving the way for mass production.

Nextron provides reliable connector solutions tailored to the four core modules of humanoid robots, ensuring stability and efficiency. Trusted and adopted by top robotics companies, Nextron supports diverse humanoid robot applications and customizes solutions to accelerate development and meet complex requirements.

References

[1] Tong, Y., Liu, H., & Zhang, Z. (2024). Advancements in humanoid robots: A comprehensive review and future prospects. *IEEE/CAA Journal of Automatica Sinica*, 11(2), 301–328. Link

[2] Brudniok, S., Albers, A., Ottnad, J., Sauter, C., & Sedchaicharn, K. (2007, June).Design of Modules and Components for Humanoid Robots. Presented at an international conference. Link