Embodied Intelligence is a cutting-edge field at the intersection of artificial intelligence and robotics, emphasizing how intelligent agents achieve autonomous learning and evolution through dynamic interaction between their bodies and the environment. Its core lies in the deep integration of perception, action, and cognition. MCU (microcontroller) serves as the key hub for robot perception and execution capabilities, receiving perceptual data from various sensors. After internal computational processing, the MCU rapidly sends instructions to actuators, driving motor operation and robotic arm movements.

Embodied intelligence also faces various security challenges, such as:

1.  User Privacy Protection In home or healthcare scenarios, robots may involve collection of privacy data such as voice interaction and facial recognition. It is necessary to ensure that user privacy is not illegally obtained.

2.  Secure Communication Protocol Support Various modules within humanoid robots (such as joint drives, sensors, and main control units) need real-time communication through bus protocols like CAN, EtherCAT, and TSN. Encryption chips can provide identity authentication and message encryption for these communication links, preventing man-in-the-middle attacks.

3.  Edge Computing and Security Coordination To achieve low-latency decision-making, some data needs to be processed locally at edge nodes (such as dexterous hands). Encryption chips combined with edge AI algorithms can complete sensitive data processing locally before encrypting and uploading, reducing cloud transmission risks.

Application Advantages:

N32H7 high-performance MCU based on Arm® Cortex®-M7 multi-core heterogeneous architecture:

·         Arm® Cortex®-M7: Performance reaches 1284 DMIPS (Cortex-M7 @600MHz)

·         Arm® Cortex®-M4: Performance reaches 375 DMIPS (Cortex-M4 @300MHz)

·         Absolute zero-wait instruction fetch: 600MHz main frequency, unlimited code size, unlimited instruction jumps, unlimited function nesting

·         Rich peripherals including EtherCAT, CAN-FD, 100ps-level ultra-high resolution timers, floating-point input/output, CORDIC, etc.

N32H47x based on Arm® Cortex®-M4F core, main frequency up to 200MHz:

·         Rich timer resources with high precision: 1 high-precision timer (125ps accuracy), 3 advanced timers (5ns accuracy), 10 general-purpose timers (5ns or 5.56ns accuracy)

·         TX and RX of 3 U(S)ARTs support full pin mapping

·         CAN-FD pins support full pin mapping

·         ADC with 4.7Msps sampling rate, supporting single-ended and differential modes

·         High-performance analog operational amplifier supporting single-ended and differential modes, GBW 40MHz, SR 20V/μs (single-ended mode) and SR 40V/μs (differential mode)

·         CORDIC supports integer and floating-point operations

Recommended Models:

·         Joint drive module: N32H765EC

·         Dexterous hand: N32H765EC / N32H47x

·         Cerebellum: N32H785EC

·         LiDAR: N32H47x / N32G430

·         IMU module: N32H47x

·         BMS: N32G45x/N32L406

·         Secure communication/Edge computing security coordination/Privacy protection: N32S032/N32S003