Integrating Intelligence Throughout Humanoid Robots: A Domestic MCU Company’s Exploration of MCU+AI (TinyML)

According to a report by Electronic Enthusiasts (Author: Wu Zipeng), when developing humanoid robot solutions, most people equate the intelligence capabilities of robots with AI large models. With high-performance AI chips paired with AI large models, the intelligence level of humanoid robots has significantly improved. However, AI large models are primarily used for high-level intelligent tasks such as semantic understanding, complex scene perception, and long-term task planning, allowing humanoid robots to perform excellently in standardized scenarios. Yet, in highly personalized scenarios that require real-time responses, AI large models cannot meet the real-time control demands. Therefore, MCU+AI (TinyML) has become an effective supplementary solution for intelligence.In the Electronic Enthusiasts’ special report on “Motor Control and Sensors in Humanoid Robots,” Mr. Zhong Xuheng, CEO of Pengpai Micro, revealed in an interview that the company has begun exploring and practicing in the MCU+AI (TinyML) field, developing an offline voice recognition + motor control single-chip solution. TinyML is an ultra-small model that originated from the development of smart sensors. TinyML significantly compresses traditional AI models, reducing computational load and storage space while maintaining model performance. Such models can be conveniently deployed on medium to small-scale MCUs. With the development of large models, they provide a quick pathway for training small models, and in the future, MCU+AI (TinyML) is expected to be widely applied in the field of humanoid robots.

Integrating Intelligence Throughout Humanoid Robots: A Domestic MCU Company's Exploration of MCU+AI (TinyML)

Mr. Zhong Xuheng, CEO of Pengpai MicroHumanoid robots require various chips, such as high-performance AI chips, memory chips, sensor chips, analog chips, and controller chips. Currently, Pengpai Micro has laid out control chips, including motor driver control chips and MCU+AI control chips that integrate AI.In the field of humanoid robots, the synergy between AI large models and MCU+TinyML is essentially a deep complement between general cognitive abilities and edge real-time responses. From a system architecture perspective, this complementarity is not traditional functional redundancy. In fact, the MCU, aided by the lightweight models optimized by TinyML, can help AI large models complete the closed loop of terminal operations, extending intelligent capabilities from the “brain” to the entire body.Clearly, MCU+AI (TinyML) is one of the concrete implementations of edge intelligence concepts in the field of humanoid robots. Mr. Zhong pointed out that in addition to using MCU+AI (TinyML) for intelligent control of motors, a full-link edge intelligence system comprising sensors + motor control + AI will also be formed at the control terminal of humanoid robots. This integration not only significantly enhances the performance of humanoid robots and increases their adaptability but also reduces costs, thereby accelerating the widespread adoption of humanoid robots. In this process, chips need to enhance AI capabilities while further exploring the deep integration of AI and chips, maximizing the capabilities of AI and chips in vertical application fields.Mr. Zhong mentioned that compared to traditional motor control, the requirements for humanoid robot motor control are much higher. For instance, the smoothness, precision, and responsiveness of motor operations in humanoid robots must meet high standards, with short response times and strong resistance to external interference. Additionally, there is an aspect that traditional motor control rarely addresses: numerous and diverse motors must maintain good synchronization to ensure that humanoid robots perform precise movements. For the coordinated control of motors, it may not rely entirely on computation but rather on training results, which fully reflects the importance of AI in the field of humanoid robots.In addition to control chips, Pengpai Micro is also focusing on the application of force sensors in humanoid robots. Force sensors play a crucial role in important components such as robotic hands and determine the application fields that humanoid robots can serve. Sensors collect physical signals and convert them into electrical signals, which the MCU can gather and perform various mathematical operations. For MCUs with AI capabilities, they can also make AI judgments, upgrading traditional sensors into true smart sensors. The addition of AI has qualitatively changed the collaboration between sensors and MCU controllers, freeing control from heavy mathematical computations, significantly improving control response speed, and reducing the energy consumption of high computational power.These innovative solutions help further explore the potential of humanoid robots, promoting their implementation in more scenarios. According to the “Humanoid Robot Industry Research Report” released by the China Humanoid Robot Industry Conference, the market size of humanoid robots in China is expected to reach approximately 5.3 billion yuan by 2025, doubling from 2024; by 2029, the market size of humanoid robots in China is expected to reach 75 billion yuan, accounting for 32.7% of the global total, ranking first in the world; by 2035, the market size is expected to exceed 300 billion yuan.Mr. Zhong stated that humanoid robots are expected to become general-purpose robots, with enormous future development potential that will greatly change various aspects of human production and life. Currently, the key to the widespread adoption of humanoid robots lies in cost-effectiveness, which requires more innovations to enhance performance while reducing production costs. Once the cost-effectiveness breaks a certain threshold, the widespread adoption and large-scale application of humanoid robots will follow. As an innovative enterprise of domestic 32-bit MCUs, Pengpai Micro will collaborate with industry partners to integrate AI technology into signal acquisition, processing, and control, innovating products to serve the industry and promote the rapid adoption of humanoid robots.

Integrating Intelligence Throughout Humanoid Robots: A Domestic MCU Company's Exploration of MCU+AI (TinyML)

Disclaimer: This article is original from Electronic Enthusiasts, please indicate the source above when reprinting. For group communication, please add WeChat elecfans999,for submission and interview requests, please email [email protected].

More Hot Articles to Read

  • Electronic Skin Breakthroughs Ignite Humanoid Robots

  • The World’s First Two-Dimensional RISC-V MPU! Named “Wuji”, from the Fudan Team

  • Revolution in Humanoid Robot Perception! Innovative Machine Vision Sensors Emerge

  • Entering the Thousand-Yuan Level! Exoskeleton Robots Sold Out, These Motor Companies Accelerate Positioning

  • Global Foundries and UMC Plan Merger! Challenging Samsung, the Global Wafer Foundry Landscape May Change

Click to Follow and Bookmark Us

Don’t miss any updates by marking us as a favorite!Integrating Intelligence Throughout Humanoid Robots: A Domestic MCU Company's Exploration of MCU+AI (TinyML)

Leave a Comment