Neural Network API for Arduino
For Arduino users, ANNHUB does not require any external API at all. In fact, ANNHUB can export a trained neural network model to Arduino format that can be compiled in Arduino IDE.
Export a trained model to Arduino format.
Choose location for trained Arduino neural network
Compile and upload a trained neural network model in Arduino IDE.
1. Define an Neural Network object
2. Perform prediction based on the neural network input
For example, in XOR parity classification, there are two inputs and 1 output
0 XOR 0 --> 0
0 XOR 1 --> 1
1 XOR 0 --> 1
1 XOR 1 --> 0
//Define Input variable
// Define Output variable
// Perform prediction
// The output should be around 1
Arduino STM32 Blue Pill
The STM32 Blue Pill is 32 bit ARM Cortex M3 micro-controller (STM32F103C8) which is low cost and compact Arduino device.
STM32 Blue Pill Arduino device
STM32 Blue Pill Arduino pinouts
- ARM Cortex M3
- 72 MHz
- 64 KB/128 KB Flash
- 20 KB RAM
- Reset button
- LED on PIN PC13
- 32 kHz Real time clock crystal
- Jump links on Boot0 and Boot1
- Micro USB connector for power and data
- ST-Link header on the top of the board.
Before STM32 Blue Pill can be used in Arduino IDE, following steps are needed to be done.
Step 1: Flashing the boot-loader
The binary of the STM32 Blue Pill boot-loader can be download in https://github.com/rogerclarkmelbourne/STM32duino-bootloader.
For Blue Pill, the "generic_boot20_pc13.bin" located in binaries folder will be used.
Step 2: Install STM32 Blue Pill Driver
The Arduino STM32 system files can be downloaded in here. After downloading, the Arduino STM32-master can be extracted into Arduino hardware folder
The STM32 driver can be installed by running "install_drivers.bat"
The Device Manager will displace STM32 device port (COM8)
Step 3: Install STM32 Blue in Arduino IDE
The Arduino IDE can be obtained from Arduino Home page.
In Arduino IDE, select Preferences (File->Preferences)
In the Additional Boards Manager URLs, add "https://github.com/stm32duino/BoardManagerFiles/raw/master/STM32/package_stm_index.json"
Go to Tools-> Board-> Boards Manager, search for "STM32F1" and install STM32 Cores that will support STM32 Blue Pill device
Select "Generic STM32F103C series" as Board; Variant: 20k RAM, 64 k Flash, and Upload method is "STM32duino bootloader"
Step 4: Test STM32 Blue in Arduino IDE
Load, compile and upload IRIS_Classification example. Successful process can be seen as below.
The STM32 Blue Pill now can classify three different IRIS flowers.