$25 Raspberry Pi add-on gets you started with edge computing AI. The RISC-V powered Grove AI HAT for Edge Computing allows developers to use neural networks for artificial intelligence projects on Raspberry Pi.
Seeed Studio is bringing RISC-V capabilities to the Raspberry Pi with the Grove AI HAT for Edge Computing, a $25 add-on to the Raspberry Pi—sitting on top, connecting using the RPi’s GPIO connector pins—equipped with a Sipeed MAIX M1 AI Module, which utilizes a Kendryte K210 processor.
The Grove AI HAT includes six Grove interfaces, including one each for Digital I/O, PWM, I2C, and UART, and two ADC, as well as an interface for a single LCD and camera each, and a microphone and accelerometer onboard. It also integrates a USB-C 2.0 connector, a JTAG & ISP UART pin header, and two 20-pin headers for I2C, UART, SPI, I2S, PWM, and GPIO.
Sipeed compares the MAIX M1 AI Module to Google’s Edge TPU accelerator, though notes that it “acts as a master controller, not an accelerator,” making it a lower-cost and lower-power solution. MAIX, according to Sipeed, “can be used for a growing number of industrial use-cases such as predictive maintenance, anomaly detection, machine vision, robotics, voice recognition, and many more. It can be used in manufacturing, on-premise, healthcare, retail, smart spaces, transportation, etc.”
The K210, the CPU at the core of the MAIX M1, is a 28nm, 64-bit dual-core RISC-V IMAFDC design, with a neutral network processor capable of object recognition.
The Grove AI HAT can be pre-ordered for $24.50, though the price will increase modestly to $28.90 after June 15, 2019, when the device goes into general availability.
The Grove AI HAT for Edge Computing(hereafter called ‘Grove AI HAT’) is built around Sipeed MAix M1 AI MODULE with Kendryte K210 processor inside. It’s a low cost but powerful raspberry pi AI hat which assists raspberry pi run the AI at the edge, it also can work independently for edge computing applications.
The MAix M1 is a powerful RISC-V 600MHz AI module features dual core 64-bit cpu, 230GMULps 16-bit KPU(Neural Network Processor), FPU(Float Point Unit) supports DP&SP, and APU(Audio Processor) supports 8 mics.
In addition to the powerful Kendryte K210 processor, the Grove AI HAT for Edge Computing board provide a wealth of peripherals: I2C/UART/SPI/I2S/PWM/GPIO. The hat also offers LCD and camera interface, which support the Sipeed 2.4inch QVGA LCD and DVP camera, it will be helpful and convenience with your AI vision project. Just like the Sipeed MAix BiT Kit for RISC-V AI+IoT, we will release the kit with camera and LCD soon. For AI voice recognition applications, we add a high-quality microphone. And for robot or motion applications, there is a onboard 3-axis accelerometers sensor, which is more accurate and easy to use compared to external sensors.
We have released varies of SIPPED AI products, we believe it is time to make it Grove, and bring all our hundreds of grove senors and grove acvotors to your AI applications. So here comes the Grove AI HAT for Edge Computing. We’ve added 6 grove connectors to this hat, including 1xDigital IO, 2xAnalog IO, 1xI2C, 1xUART and 1xPWM. On top of that, based on kendryte-standalone-sdk, we added the full ArduinoCore-API interface to support Arduino IDE, Linux, Windows, Mac OS X and other development environments. Which means you can run Grove Arduino Libraries and many excellent Arduino libraries on this board easily.
We hope this board may help you with your edge computing, AI vision, voice recognition, and other AI projects, just enjoy it.
Why does RISC-V matter? $25 Raspberry Pi add-on gets you started with edge computing AI
RISC-V is a freely-available instruction set architecture (ISA) made available under the BSD license, and requires no patent royalties for manufacture or implementation. Any group that wishes to implement or extend RISC-V in commercial products can do so, without being required to disclose their changes to the public. This makes it particularly appealing for commercial use in embedded devices, as manufacturers can save on licensing fees associated with ARM or MIPS ISAs—both of which are fundamentally RISC in principle.
RISC-V has a great deal of momentum behind it, with the Linux Foundation and the RISC-V Foundation partnering with Google and Western Digital to found the CHIPS Alliance to promote the adoption and development of RISC-V implementations. Western Digital, which has publicly committed to “shipping two billion RISC-V cores annually” on products the company manufactures, is opening its SweRV RISC-V implementation and associated tools.
For more on RISC-V, learn how to start developing for RISC-V with the $49, Arduino-compatible HiFive1, and learn about the high-performance, Linux-capable Hi-Five Unleashed single-board computer.
Features – $25 Raspberry Pi add-on gets you started with edge computing AI
- Processor: Sipeed MAIX-I module w/o WiFi ( 1st RISC-V 64 AI Module, K210 inside )
- 1x USB 2.0 Device, Type C(Power and Programming)
- 6x Grove Interface: include 1x Digital IO, 1x PWM, 1x I2C, 1x UART, 2x ADC
- 1x Power LED, 1x Boot LED
- 1x Reset Button, 1x Boot Button
- 1x LCD Interface
- 1x Camera Interface
- 1x Digital Mic
- 1x Accelerometers Sensor
- 1x JTAG & ISP UART Pin Header
- 2x 20 Pin Header with I2C, UART, SPI, I2S, PWM, GPIO
Specification $25 Raspberry Pi add-on gets you started with edge computing AI
|CPU||K210 RISC-V Dual Core 64bit, 400Mh(Max. 600Mhz)|
|FPU||IEEE754-2008 compliant high-performance pipelined FPU|
|KPU(Neural Network Processor )||• Supports the fixed-point model that the mainstream training framework trains according to specific restriction rules
• Support for 1×1 and 3×3 convolution kernels
• Support for any form of activation function
• The maximum supported neural network parameter size for real-time work is 5MiB to 5.9MiB
|APU(Audio Processor)||• Up to 192kHz sample rate
• Up to 8 channels of audio input data, ie 4 stereo channels
|Debugging Support||High-speed UART and JTAG interface for debugging|
can not exceed 5.5V
|Operating Ambient Temperature||-20 – 70℃|
|ADC||Onboard 16-bit ADC|
|3-Axis Accelerometers||ADXL345 ±2 g/±4 g/±8 g/±16 g|
|Micrephone||Sensitivity:-26 dB;SNR:61 dB|
|GPIO||2x 20 Pin Header compatible with Raspberry Pi|
|Grove Interface||1x Digital IO, 1x PWM, 1x I2C, 1x UART, 2x ADC|
|USB Interface||USB 2.0 Device, Type C(Power and Programming)|
|LCD Interface||8bit MCU LCD 24P 0.5mm FPC connector|
|DVP Camera Interface||24P 0.5mm FPC connector|
|Button||1x Reset Button, 1x Boot Button (can be used as User Button)|
|LED||1x Power LED, 1x Boot LED(can be used as User LED)|
- All digital and analog IO interface levels are 3.3V. Please do not input more than 3.3V, otherwise the CPU may be damaged.
- The input power supply voltage is 5V and cannot exceed 5.5V.
- AI for Edge Computing
- Smart Building
- Medical equipment
- Automation & Process Control