Robot components are rapidly evolving towards high performance, low cost, modularity, and domestic production, with significant breakthroughs in core technologies. As humanoid robots move from the laboratory to large-scale deployment, the focus of the industry chain is shifting upstream to core components, driving simultaneous improvements in performance and cost reductions.
1. Reducers: Accelerated Domestic Substitution, Leap in Precision and Consistency
The domestic production rate of harmonic reducers has exceeded 50%, with product lifespan and precision reaching international first-tier levels.
The core manufacturing process of RV reducers has been mastered, significantly improving consistency in mass production and breaking the long-standing reliance on imports.
New planetary roller screws are gradually replacing traditional ball screws in high-load, long-life applications, becoming key components for linear drives.
2. Servo Systems: Dual-Wheel Drive of Integration and Self-Developed Algorithms
Domestic servo motors have become readily available, with continuous optimization in response speed, efficiency, and torque density.
The self-developed FOC control algorithm is implemented entirely in hardware using FPGA, achieving a repeatability accuracy of 0.01 degrees, supporting high-dynamic operation requirements.
The combination of frameless torque motors and coreless motors is widely used, meeting the dual requirements of compact joint layout and high explosive power output.
3. Sensors: Multimodal Fusion, Tactile and Force Sensing Become New Focuses
The localization of six-dimensional force sensors is accelerating, with Bluedot Touch achieving a market share of 72.6%, breaking the monopoly of the US, Japan, and Germany.
A humanoid robot requires 4 six-dimensional force sensors + 28 joint torque sensors, with the force control system accounting for approximately 15% of the total BOM cost.
Tactile sensors (electronic skin) are entering a period of rapid development, with capacitive and magnetoelectric Hall effects being the current mainstream technologies.
Visual sensors are evolving towards 3D, multimodal, and integrated sensing and computing, with new technologies such as ToF, structured light, and event sensors being widely applied.
