Intel × BridgeDP: Upgrading and Innovating Humanoid Robot Motion Control Technology
BridgeDP's deep technical expertise in robot motion control algorithms, combined with Intel's powerful computing performance and advanced computing platforms, will bring new breakthroughs to robot motion control technology

The humanoid robot industry has achieved explosive growth in a short period and has become a core new arena of global technological competition. However, the humanoid robot industry as a whole is still in its early development stage, and there are still many technical shortcomings. Among the key technologies that humanoid robots face, motion control serves as the robot's "cerebellum," controlling every movement and directly determining the smoothness and balance of its actions.
At this critical stage where industry pain points and opportunities coexist, Intel has partnered with BridgeDP, a company focused on general-purpose robot motion control technology. The two parties fully leverage their complementary technical strengths: Intel provides hardware support with its x86 architecture computing foundation, while BridgeDP optimizes software with its expertise in motion control algorithms, jointly creating an integrated "computing power + algorithms" solution. This provides a new path to break through the technical bottleneck of humanoid robot motion control and helps the industry accelerate from "technological breakthroughs" to "scenario deployment."
BridgeDP: Focusing on Robot Motion Control with Multi-Algorithm Integration for Customized Control Solutions
Since its establishment in 2023, BridgeDP has been committed to solving key technical challenges in the motion control of humanoid, quadruped, and wheel-legged robots, providing them with precise, flexible, and efficient motion control solutions. To meet the motion control needs of various robots, BridgeDP conducts tests based on the x86 architecture and deploys motion control solutions by invoking basic motion control modules as needed through technologies such as Sim2Real training, reinforcement learning, and deep imitation.
The omnidirectional locomotion control solution developed by BridgeDP achieves a balance between "realism" and "robustness" through an innovative algorithmic architecture. The solution can be adapted to robots of different configurations, including humanoid, quadruped, and wheel-legged, and can learn various locomotion styles. It supports not only basic movements such as forward, backward, lateral translation, turning in place, walking, and running, but also complex scene operations such as climbing stairs with perception.
The full-body imitation motion control solution developed based on algorithms such as DeepMimic and AMP enables high-precision transfer of complex human movements to humanoid robots through a two-stage "imitation learning + reinforcement learning" approach. It can support complex movements such as dancing and martial arts, as well as customized motions. As of the third quarter of 2025, its motion control solutions have been deployed in nearly 50 robot manufacturers.


The image above shows a quadruped robot dog overcoming obstacles with perception.
The image below shows full-body motion control of a humanoid robot.
During testing and deployment, BridgeDP believes that in motion control scenarios,the Intel N97 processor offers strong hardware performance adaptability, compact size, and sufficient performance. It can stably meet the operational requirement of a 1000Hz control cycle during motion control, ensuring the smooth execution of high-frequency control commands. At the same time, relying on the more complete software ecosystem of the x86 architecture, it simplifies the actual deployment process of motion control, reduces debugging difficulty, and improves the efficiency of technology implementation and iteration.
In AI + motion control fusion scenarios,the Intel® Core™ Ultra 200H series processors have strong multitasking capabilities and ample performance headroom. While simultaneously running AI inference and motion control and ensuring real-time performance for both, they can also run performance analysis programs in parallel, breaking through the real-time limitations of the scenario and improving the stability of multi-task collaboration.
Intel: Deepening Robot Control Technology and Leading Innovation in Cerebrum-Cerebellum Fusion Architecture
The Intel® Core™ Ultra 200H series processors (codenamed Arrow Lake-H) are based on an updated manufacturing process and adopt an innovative heterogeneous computing architecture, integrating high-performance CPU cores, Arc GPU, and a dedicated NPU to deliver powerful AI computing of 96 TOPS. This processor achieves a groundbreaking "cerebrum-cerebellum fusion" architecture in the field of robot control: high-frequency CPU cores perform the "cerebellum" function, executing deterministic real-time motion control tasks such as trajectory planning and multi-axis coordinated control; the NPU and GPU play the role of the "cerebrum," handling cognitive intelligent decision-making such as environmental perception, task understanding, and path planning. Through multi-core parallel processing and heterogeneous collaborative computing technology, the processor achieves deep integration of millisecond-level precision motion control and complex AI inference on a single hardware platform, providing a complete integrated cerebrum-cerebellum computing solution for embodied AI robots.

BridgeDP's control frequency and time consumption test based on 255H
(The green curves on the left, from top to bottom, are the operating frequencies (Hz) of the read, update, and write modules)
(The yellow curves on the right, from top to bottom, are the time consumption (s) of the read, update, and write modules)
(The red curve in the lower right corner is the total time consumption of read, update, and write)
*Whether read, update, or write, the execution speed is very fast and stable.
The Intel® N97 processor adopts an advanced process technology, integrates 4 CPU cores with a maximum turbo frequency of 3.6GHz, and includes Intel UHD Graphics. With only 12W low power consumption, it specializes in the "motion control cerebellum" function in robot applications. Thanks to its optimized architecture design and excellent real-time responsiveness, this processor can precisely execute deterministic motion control tasks such as trajectory control, multi-axis synchronization, and servo feedback processing. Its integrated graphics can also handle real-time data visualization and human-machine interface display. The N97 processor achieves excellent energy efficiency and power control while ensuring high-performance motion control processing, with good temperature adaptability and long-term stability, and supports rich I/O interface expansion. It provides an ideal "cerebellum" computing platform for mobile robots, collaborative robots, small humanoid robots, and other power-sensitive applications requiring precise motion control.

BridgeDP's control frequency and time consumption test based on N97
(The green curves on the left, from top to bottom, are the operating frequencies (Hz) of the read, update, and write modules)
(The yellow curves on the right, from top to bottom, are the time consumption (s) of the read, update, and write modules)
(The red curve in the lower right corner is the total time consumption of read, update, and write)
*Operation is very stable and can meet the real-time control requirement of 1000Hz.

Comparison with test results of another architecture platform
(The green curves on the left, from top to bottom, are the operating frequencies (Hz) of the read, update, and write modules)
(The yellow curves on the right, from top to bottom, are the time consumption (s) of the read, update, and write modules)
(The red curve in the lower right corner is the total time consumption of read, update, and write)
*This figure obtains information through a network cable connection. As shown, the jitter amplitude of information transmission is relatively large.
Partnership Outlook: Computing Power Optimizes Algorithms to Create Efficient and Intelligent Motion Control Solutions
The cooperation between BridgeDP and Intel is an effective exploration of complementary strengths. BridgeDP's deep technical expertise in robot motion control algorithms, combined with Intel's powerful computing performance and advanced computing platforms, will bring new breakthroughs to robot motion control technology.
In terms of product development, the cooperation will aim to create more efficient and intelligent robot motion control solutions. By using Intel's computing power to optimize BridgeDP's motion control algorithms, robots will achieve more precise motion control and higher stability during movement. In terms of application expansion, the cooperation will promote the widespread application of robot motion control technology in more fields, driving robot application scenarios deeply into logistics distribution, medical rehabilitation, home services, and other areas. In the future, with continuous technological iteration and ecosystem improvement, humanoid robots will truly possess the core capabilities of precise motion, intelligent interaction, and autonomous decision-making, becoming a key force in promoting industrial intelligence and upgrading home services.





界面新闻
深视新闻
深圳特区报
财新
36氪
钛媒体
中国新闻周刊
蓝鲸新闻
Z Finance
投资家网
雷峰网
亿欧
环球网
南方日报
极客公园



