原标题：Maxim MAX78000超低功耗人工智能(AI) MCU开发方案

　　Maxim公司的MAX78000是集成了基于FPUMCU和卷积神经网络加速器的超低功耗人工智能(AI) 处理器.它是新品种的AI 微控制器,用来建立神经网络,以实现超低功耗和

　　在物联网(IoT)的边缘现场直播.该器件组合了最有能量效率的AI处理和现已验证了超低功耗MCU.基于硬件的卷积神经网络(CNN)加速器是电池供电的应用能实现人工智能推断而仅需化费微焦尔的能量. MAX78000是先进的系统级芯片(SoC),具有FPU CPU的ArmR CortexR-M4核,以及超低功耗深度神经网络加速器.CNN引擎具有加权存储器442KB,支持1,2,4和8位权重.CNN加权存储器是基于SRAM的,这样AI网络升级能即时进行.CNN引擎还有512KB数据存储器,CNN架构是高度灵活的,允许网络用通常的工具包如PyTorch和 TensorFlowR进行培训,然后转换成可以由MAX78000采用已验证过的工具来执行.带FPU的Arm Cortex-M4处理器工作高达100MHz,具有16KB指令缓冲的最佳性能,用于SRAM的可选择误差修正码(ECC-SEC-DED),而332位RISC-V协处理器工厂作频率高达60MHz主要用在目标检测和分类,音频处理如多个关键字识别,声音分类,噪音消除,面部识别以及时间系列数据处理如心率/健康信号分析,多传感器分析以及预测性维修.本文介绍了MAX78000主要优势和特性,简化框图,时钟方案框图以及评估板MAX78000 EVK优势和特性,电路图和材料清单.

　　Artificial intelligence (AI) requires extreme computational horsepower, but Maxim is cutting the power cord from AI insights. The MAX78000 is a new breed of AI microcontroller built to enable neural networks to execute at ultra-low power and live at the edge of the IoT. This product combines the most energy-efficient AI processing with Maxims proven ultra-low power microcontrollers. Our hardware-based convolutional neural network (CNN) accelerator enables battery-powered applications to execute AI inferences while spending only microjoules of energy.

　　The MAX78000 is an advanced system-on-chip featuring an ArmR CortexR-M4 with FPU CPU for efficient system control with an ultra-low-power deep neural network accelerator.

　　The CNN engine has a weight storage memory of 442KB, and can support 1-, 2-, 4-, and 8-bit weights (supporting networks of up to 3.5 million weights). The CNN weight memory is SRAM-based, so AI network updates can be made on the fly. The CNN engine also has 512KB of data memory. The CNN architecture is highly flexible, allowing networks to be trained in conventional toolsets like PyTorch and TensorFlowR, then converted for execution on the MAX78000 using tools provided by Maxim.

　　In addition to the memory in the CNN engine, the MAX78000 has large on-chip system memory for the microcontroller core, with 512KB flash and up to 128KB SRAM. Multiple high-speed and low-power communications interfaces are supported, including I2S and a parallel camera interface (PCIF).

　　The device is available in 81-pin CTBGA (8mm x 8mm, 0.8mm pitch) and 130-pin WLP (4.6mm x 3.7mm, 0.35mm pitch) packages.

　　MAX78000主要优势和特性:

　　● Dual Core Ultra-Low-Power Microcontroller

　　• Arm Cortex-M4 Processor with FPU Up to 100MHz

　　• 512KB Flash and 128KB SRAM

　　• Optimized Performance with 16KB Instruction Cache

　　• Optional Error Correction Code (ECC-SEC-DED) for SRAM

　　• 32-Bit RISC-V Coprocessor up to 60MHz

　　• Up to 52 General-Purpose I/O Pins

　　• 12-Bit Parallel Camera Interface

　　• One I2S Master/Slave for Digital Audio Interface

　　● Neural Network Accelerator

　　• Highly Optimaized for Deep Convolutional Neural Networks

　　• 442k 8bit Weight Capacity with 1,2,4,8-bit Weights

　　• Programmable Input Image Size up to 1024 x 1024 pixels

　　• Programmable Network Depth up to 64 Layers

　　• Programmable per Layer Network Channel Widths up to 1024 Channels

　　• 1 and 2 Dimensional Convolution Processing

　　• Streaming Mode

　　• Flexibility to Support Other Network Types, Including MLP and Recurrent Neural Networks

　　● Power Management Maximizes Operating Time for Battery Applications

　　• Integrated Single-Inductor Multiple-Output (SIMO) Switch-Mode Power Supply (SMPS)

　　• 2.0V to 3.6V SIMO Supply Voltage Range

　　• Dynamic Voltage Scaling Minimizes Active Core Power Consumption

　　• 22.2μA/MHz While Loop Execution at 3.0V from Cache (CM4 only)

　　• Selectable SRAM Retention in Low-Power Modes with Real-Time Clock (RTC) Enabled

　　● Security and Integrity

　　• Available Secure Boot

　　• AES 128/192/256 Hardware Acceleration Engine

　　• True Random Number Generator (TRNG) Seed Generator

　　MAX78000应用:

　　● Object Detection and Classification

　　● Audio Processing: Multi-Keyword Recognition, Sound Classification, Noise Cancellation

　　● Facial Recognition

　　● Time-Series Data Processing: Heart Rate/Health Signal Analysis, Multi-Sensor Analysis, Predictive Maintenance

　　图1. MAX78000简化框图

　　图2. MAX78000时钟方案框图

　　评估板MAX78000 EVK

　　The MAX78000 evaluation kit (EV kit) provides a platform for leveraging the capabilities of the MAX78000 to build new generations of artificial intelligence (AI) devices.

　　The EV kit features a socketed MAX78000. Onboard hardware includes a digital microphone, a gyroscope/ accelerometer, parallel camera module support and a 3.5in touch-enabled color TFT display. A secondary display is driven by a power accumulator for tracking device power consumption over time. Uncommitted GPIO as well as analog inputs are readily accessible through 0.1in pin headers. Primary system power as well as UART access is provided by a USB Micro-B connector. A USB to SPI bridge provides rapid access to onboard memory, allowing large networks or images to load quickly.

　　评估板MAX78000 EVK优势和特性:

　　●● Power Accumulator with Dedicated Display to Track Device Power over Time

　　●● Onboard Digital Microphone

　　●● Onboard Accelerometer/Gyroscope

　　●● SWD JTAG 10-Pin Header

　　●● RISC-V Coprocessor JTAG 10-Pin Header

　　●● 16M-Byte QSPI Flash

　　●● Select GPIOs Accessible through 0.1in Headers

　　●● Four ADC Inputs with Optional AA Filter

　　●● Touch-Enabled, 3.5in, 320 x 240 Color TFT Display

　　●● UART Access through USB Bridge

　　●● QSPI Memory Access through USB Bridge

　　●● All IC Power Rails May Be Isolated by Jumpers for Individual Current Measurements

　　●● Two General-Purpose LEDs and Two General- Purpose Pushbutton Switches

　　评估板MAX78000 EVK包含:

　　●● EV Kit Board with Socketed MAX78000

　　●● MAX32625PICO Debugger with Cables

　　●● Olimex ARM-USB-OCD-H

　　●● Olimex ARM-JTAG 20-10 Adapter

　　●● Camera Module

　　●● 2 USB Standard-A to USB Micro-B Cables

　　●● 1 USB Standard-A to USB Standard-B Cable

　　●● Extra Shunts

　　图3.评估板MAX78000 EVK外形图

　　图4.评估板MAX78000 EVK电路图(1)