NXP delivers new innovations for advanced physical AI with NVIDIA
NXP has announced robotics solutions for real-time data processing, data transport, and networking designed to support sensor fusion, machine vision, and motor control. The solutions were developed in collaboration with NVIDIA and implement NVIDIA Holoscan Sensor Bridge with NXP system-on-chip devices, reducing the number of discrete components and lowering footprint, power consumption, and system cost while simplifying software for robotic sensing and actuation, including humanoid robot designs.
Humanoid and other advanced robotic systems require low-latency data processing and communication to support synchronised motion, multiple sensors, and distributed actuation. These systems depend on reliable data transport between sensors, processors, and control units across the robot body.
NXP’s integrated robot body solutions combine edge processing with low-latency networking to support real-time communication. The solutions integrate NVIDIA Holoscan Sensor Bridge with NXP software support, enabling real-time processing and direct data transport between sensors, control nodes, and central processing units, reducing latency in robotic control systems.
The platform combines NXP edge processors, motor-control microcontrollers, networking devices, and high-speed data transport technology with NVIDIA AI processing infrastructure to form a system architecture intended for humanoid and other complex robotic platforms.
Initial Holoscan Sensor Bridge-ready solutions include a machine-vision platform based on the i.MX 95 applications processor for high-bandwidth sensor data, and a motor-control platform based on i.MX RT1180 crossover MCUs connected through an S32J TSN switch. The motor-control design supports industrial networking protocols including EtherCAT and TSN. The solutions are designed to reduce component count, power consumption, and system size while supporting safety, security, and real-time performance requirements in robotic systems.
