According to reports from Electronic Enthusiasts (by Wu Zipeng), with continuous breakthroughs in core technologies such as AI large models, motor drives, and multimodal sensing, humanoid robots are at the critical point of technological breakthroughs and large-scale production, showcasing tremendous development potential. The “Humanoid Robot Industry Research Report” indicates that the market size for humanoid robots in China is expected to reach 75 billion yuan by 2029, and is projected to exceed 300 billion yuan by 2035.In the Electronic Enthusiasts’ special feature on “Motor Control and Sensors for Humanoid Robots,” Zhao Xiangyuan, the Technical Support Director for Texas Instruments in China, stated that robots are actively improving people’s daily lives in both industrial and personal domains. As a product of the integration of artificial intelligence and robotics technology, humanoid robots are rapidly penetrating multiple application scenarios, especially showing great potential in caregiving and household areas. For example:· It is estimated that two-thirds of the elderly cannot independently meet their basic living needs and require some form of support or care. Humanoid robots can assist healthcare personnel in daily caregiving tasks, such as helping immobile patients move, providing companionship services, and even monitoring patients’ health conditions, thereby reducing the workload of caregivers and improving the quality of care.· In household scenarios, humanoid robots can take on household chores such as cooking, cleaning, and organizing, providing users with a more convenient and intelligent lifestyle.
Zhao Xiangyuan, Technical Support Director for Texas Instruments in ChinaFacing this historic opportunity for humanoid robots, how will Texas Instruments (TI), as a global leading semiconductor company, respond? It is believed that this article will provide a deeper understanding.
The Chip Demand for Humanoid Robots
As the “brain” and “nerve center” of humanoid robots, chips play a fundamental and decisive role in their design and implementation. The types of chips involved include AI master control chips, motion control chips, communication chips, etc.Zhao Xiangyuan believes that in the field of humanoid robots, chip demand mainly focuses on control precision, monitoring and safety scanning, flexibility, and functional safety. Specifically, this is reflected in:
- Precision: Humanoid robots need to perform complex actions and tasks, which places extremely high demands on motion control precision, making high-performance control chips crucial.
- Monitoring: Robots need to monitor their surrounding environment in real-time and conduct safety scans during operation to avoid collisions or accidents, which poses challenges to the chip’s real-time processing capability and sensor integration capability.
- Flexibility: Humanoid robots must adapt to diverse scenarios and tasks, requiring chips to possess high flexibility and scalability to support different algorithms and functional requirements.
-
Functional Safety: Functional safety is core to robot design, and chips must meet strict international safety standards to ensure the reliability and safety of robots in various environments.
Texas Instruments mentioned in its technical documentation on “Motor Control in Humanoid Robots” that currently, although relevant standards have been established for collaborative robots and industrial robots, there are no standards specifying functional safety requirements for humanoid robots. As demand continues to grow, it is expected that standard organizations will set safety requirements for humanoid robots in the future. Before these safety requirements are established, humanoid robot designers must conduct relevant research on current system designs to minimize the workload of redesign in the future. Standards such as ISO13482, ISO10218, and ISO 3691-4 can clarify future expectations to some extent.
The Specificity of Chip Demand for Humanoid Robots
From Zhao Xiangyuan’s description, it is evident that the chip demand for humanoid robots significantly differs from that of past industrial and consumer robots. In practical applications, this difference may be more intuitively reflected, such as in motor drive and control.Designing drivers for humanoid robots requires precision, flexibility, and innovation. Compared to traditional motor control, the motor control demands for humanoid robots are more complex and diverse. As the structure of humanoid robots becomes increasingly complex and precise, with degrees of freedom (DOF) continuously increasing, diverse requirements arise for drivers located at different positions (such as hands, neck, shoulders, elbows, waist, hips, knees, and ankles), necessitating comprehensive breakthroughs in high-precision control, real-time response, multi-axis coordination, energy efficiency optimization, and safety. For example, in caregiving scenarios, robots need to perform grasping or assisting actions with millimeter-level precision; in household scenarios, robots need to quickly adjust their posture in dynamic environments to avoid collisions. These demands pose higher challenges to motor control technology.To address these challenges, humanoid robot designers need to specifically optimize the characteristic performance of motor control solutions, including communication interface architecture, position sensing, motor types, motor control algorithms, power stage requirements, electronic circuit size, and functional safety. For instance, given the distribution of drivers within the robot, optimizing communication with all drivers while minimizing wiring is particularly important. There are various methods for achieving optimization, with daisy chain communication and linear bus topology being the most commonly used.
Texas Instruments’ Layout in Humanoid Robot Motor Drives
In response to the specific chip demands of humanoid robots (referred to as humanoid robots by Texas Instruments), the company offers a comprehensive range of integrated circuit products that enable engineers to meet various design specifications, thereby building robots that can smoothly interact with their environments.Zhao Xiangyuan mentioned that building humanoid robots based on Texas Instruments’ devices and solutions has three significant advantages:
- Intelligent, Safe, and Compact Robot Motor Drivers: Utilizing Texas Instruments’ GaN devices and gate drivers, it is possible to construct small, efficient, and safe servo drive systems. High-precision current detection and encoders enable more accurate motor control.
- Intelligent Detection, Perception, and Navigation: Texas Instruments’ processor series can provide the high-performance edge AI computing required for perception, sensor fusion, and navigation. Texas Instruments’ millimeter-wave radar product series can achieve high-precision detection functions for humanoid robots, enhancing their environmental perception and autonomous navigation capabilities.
-
Functional Safety Requirements: By integrating on-chip safety features in Texas Instruments’ analog and embedded integrated circuits, safety standards such as IEC 61508, ISO 13849, and ISO 26262 can be met at a lower system-level cost, ensuring the safety and reliability of robots in various application scenarios.
From the technical documentation on “Motor Control in Humanoid Robots,” it is evident that Texas Instruments has a rich selection of products available for this application. Below, we will introduce several typical chips.First is the F29H85x, a real-time microcontroller designed for high-performance motor control, built on a new C29x core introduced by Texas Instruments last year. This product significantly enhances the capabilities of motor control systems in the field of humanoid robots. Its key features include: a main frequency of up to 800 MHz and strong floating-point computation capabilities, enabling microsecond-level real-time control; specifically designed to meet the latest automotive and industrial safety standards, with integrity levels reaching ASIL-D and SIL-3; an isolated hardware security module (HSM) protects the system from unauthorized access and cyber threats; functional safety and information security units (SSU) can avoid interference between code threads without affecting real-time performance.The outstanding product performance of the F29H85x allows humanoid robots to perform complex actions with extremely high precision. Its hardware diagnostics and error detection capabilities ensure the reliability and safety of the motor control system in complex environments; it supports multi-axis motor control, capable of simultaneously handling motion control tasks for multiple joints, such as maintaining balance and flexibility while walking or grasping objects; efficient energy management and low-power design significantly extend the robot’s operating time, making it more suitable for long-duration applications; a rich array of interfaces and high-precision analog peripherals facilitate seamless connections with sensors, drivers, and other subsystems.Next is the AM261x microcontroller, designed to meet the complex real-time processing and control needs of next-generation industrial and automotive embedded projects. The AM261x uniquely combines advanced computing with excellent real-time control peripherals to meet the growing performance demands of applications such as hybrid/electric vehicles (on-board chargers, DC/DC converters, battery management systems), two-axis servo drivers, industrial digital power control (energy storage, string inverters), and other general real-time constrained systems, all of which are critical applications in humanoid robot design.The R5F core in the AM261x microcontroller is arranged in clusters and features 256KB of shared tightly coupled memory (TCM) and 1.5MB of shared SRAM. For different functional safety configurations, the Arm® core can be programmed to operate using lock-step options. The on-chip memory, peripherals, and interconnects include extensive ECC, enhancing reliability. In addition to the fine-grained firewall managed by the HSM, the AM261x device also provides encryption acceleration and secure boot features for developers to design more secure systems.Zhao Xiangyuan stated that Texas Instruments’ devices fully consider redundancy and functional safety compliance, and offer various smaller package options that help optimize board space, making them better suited for joint parts of collaborative robots and humanoid robots. Additionally, Texas Instruments’ GaN devices integrate gate drivers, achieving dual optimization of size and efficiency for robots. This not only helps manufacturers reduce space occupancy and optimize power consumption but also maximizes the performance of robots, thereby extending operating time and enhancing load capacity.In addition to high-performance chips, Texas Instruments also provides reference designs that can enhance developers’ design efficiency. For humanoid robot motor driver design, the company offers a 48V/16A compact three-phase GaN inverter reference design TIDA-010936 for integrated motor drivers, as well as a 48V, 4kW compact three-phase inverter reference design TIDA-010956.TIDA-010936 showcases a high power density 12V to 60V three-phase power stage using three 100V, 35A GaN half-bridges with integrated GaN FETs, drivers, and bootstrap diodes, specifically designed for motor-integrated servo drivers and robotic applications. By using the IN241A current detection amplifier, precise phase current detection is achieved; at the same time, DC link voltage and phase voltage are measured, enabling verification of advanced sensorless designs such as InstaSPIN-FOC™. This design provides a 3.3V I/O interface compatible with Texas Instruments’ BoosterPack for quick and easy evaluation of Texas Instruments’ GaN technology. Key features of TIDA-010936 include:
- High efficiency (peak at 99.3%) at 40kHz PWM, supporting stable operation at an ambient temperature of 25°C with continuous currents up to 16Arms without the need for a heatsink.
- Compact GaN half-bridge power stage achieves high power density and simplifies PCB layout.
- GaN half-bridge supports operation at higher PWM frequencies, helping to reduce the height of DC bus capacitors while using ceramic capacitors instead of electrolytic capacitors.
- Zero reverse recovery loss reduces switching node oscillation.
-
16.6ns low dead time maximizes phase voltage distortion reduction. Using a 1mΩ shunt and a current detection amplifier with high PWM suppression capability, precise detection of phase currents in the ±33A range is achieved.
TIDA-010956 showcases a three-phase inverter with a nominal DC input voltage of 48V and a rated output current of 85ARMS. The 100V smart half-bridge gate driver DRV8162L enables a compact, robust, and efficient power stage. It is recommended to implement a safety torque-off (STO) function using a dual power architecture with multiple channel shutdown paths of the RV8162L. With the internal VDS monitoring and protection features of DRV8162L, the power stage is protected against breakdown overcurrent faults or output short circuits. Using INA241A, precise phase current detection is achieved. This design provides a 3.3V I/O interface for quick and easy evaluation with host controllers such as the C2000™ MCU. Key features of TIDA-010956 include:
- A three-phase inverter with 24V DC to 60V DC input and 85ARMS continuous output current.
- DRV8162L features an intelligent half-bridge gate driver with on-chip hardware protection.
-
Fully protected power stage: breakdown, overcurrent, short circuit, undervoltage, and overheat protection.
Zhao Xiangyuan stated that Texas Instruments collaborates with engineers and original equipment manufacturers to address functional safety compliance challenges while also focusing on developing solutions for logistics and automation. At the system level, Texas Instruments directly confronts complex design challenges, providing multi-chip or single-chip solutions. Texas Instruments is also pioneering practices that allow various robots to work alongside employees, helping to enhance production efficiency. Meanwhile, Texas Instruments leverages these practical experiences to continuously refine product quality, laying a solid foundation for future development.
Texas Instruments’ Layout in Intelligent Sensing for Humanoid Robots
In terms of perception, humanoid robots need to sense the environment and gather data through various sensors to achieve functions such as navigation, interaction, and operation. Among them, millimeter-wave radar can be used to detect the distance, speed, and angle of targets, suitable for applications such as obstacle avoidance, human detection, and gesture recognition in complex environments.Zhao Xiangyuan pointed out that Texas Instruments’ industrial millimeter-wave sensors not only provide distance information but also relative speed information of on-site personnel, allowing the robot to decelerate or stop based on proximity and speed. These radar sensors can operate in harsh environments such as rain, smoke, low light, and dusty factory conditions, providing more value compared to vision-based and laser radar sensors. Their main features include:
- IEC 61508 certified, achieving up to SIL 2 level.
- 60GHz and 77GHz frequency bands provide scalable deployment in different regions.
- FMCW radar measures the position and speed of objects.
-
3D detection with a wide azimuth coverage of up to 130º.
At the same time, Zhao Xiangyuan emphasized that sensor fusion is crucial for humanoid robots for several reasons: First, using only one type of sensor in humanoid robots has obvious limitations, including incomplete or inaccurate data collection. For example, cameras may struggle with depth perception, low light, or detecting non-visual elements, while laser radar sensors may yield inaccurate readings due to laser bouncing up and down as the robot moves. Data collection limitations may lead to errors in navigation, object manipulation, and environmental interaction. Second, sensor fusion technology can integrate data from multiple sensors, providing a more accurate, reliable, and comprehensive understanding of the robot’s environment, thus addressing the aforementioned issues. Furthermore, by combining inputs from various sensing modalities, humanoid robots can make more informed decisions, enhancing their ability to perform complex tasks, such as navigating uneven terrain, grasping objects of different shapes and sizes, and interacting in dynamic real-world environments.Texas Instruments is also committed to providing sensor fusion solutions, such as camera and radar sensor modules, as well as the IMX219 camera and IWR6843ISK EVM millimeter-wave radar sensor. This module implements an object-level fusion approach that applies a camera-vision processing chain focused on object clustering and tracking, along with a radar processing chain, allowing users to track and detect objects in a three-dimensional environment. Users have demonstrated the many potentials and functionalities of sensor fusion through the robot SDK.Zhao Xiangyuan stated in an interview that as technology matures and costs gradually decrease, humanoid robots will expand from traditional fields such as industry, caregiving, and households to emerging fields such as education, entertainment, and retail, becoming intelligent assistants across multiple industries. The deep integration of technologies such as artificial intelligence, 5G, and the Internet of Things will drive humanoid robots towards a more intelligent and interconnected direction, enabling them to better understand and adapt to complex environments.Texas Instruments deeply recognizes the continuous and far-reaching impact of robotic technology and is fully committed to leveraging semiconductors to drive the realization of this vision. In the field of robotic technology, Texas Instruments is dedicated to developing cost-effective, high-performance devices to flexibly respond to a diverse range of tasks from simple to complex, providing solid support for enterprises and helping to create innovative and outstanding end products. To overcome the various challenges that current and future robots may face in practical applications, Texas Instruments has partnered with multiple companies to jointly develop sensor and motor control solutions.
Disclaimer: This article is original from Electronic Enthusiasts, please indicate the source above when reprinting. For group communication, please add WeChat elecfans999,for submission of interview requests, please send an email to [email protected].
More Hot Articles to Read
-
Electronic Skin Breaks Out, Igniting Humanoid Robots
-
The World’s First Two-Dimensional RISC-V MPU! Named “Wuji”, from the Fudan Team
-
Humanoid Robot Perception Revolution! 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! Impacting Samsung, the Global Wafer Foundry Landscape May Change
Click to Follow and Star Us
Set us as a star to not miss any updates!