This package contains a demo project for the STM32F407 Discovery kit using Picovoice platform.
For this demo, you need to:
- Download and install STM32CubeIDE, which is an all-in-one multi-OS development tool for STM32 microcontrollers.
- Follow steps mentioned in readme for STM32Cube middleware for audio PDM to PCM conversion.
In order to compile and run the demo project on a STM32F407 discovery board, perform the following steps:
- Open STM32CubeIDE
- Click
File
>Open Projects from file system...
to display theImport Projects
dialog box. Select the stm32f407g-disc1 folder from this repository, and then press theFinish
button. - Copy the
Inc
andLib
folders from the downloaded PCM2PDM library to /Middlewares/ST/STM32_Audio/Addons/PDM - Click
Project
>Build All
- Connect the board to the computer and press
Run
>Debug
.
⚠️ printf()
uses the SWO connector and the trace port 0. For more information, refer to STM32 microcontroller debug toolbox, Chapter 7.
For this demo, the default wake word is Picovoice
and the context is Smart Lighting
. The engine can recognize commands such as
Picovoice, turn off the lights.
or
Picovoice, set the lights in the bedroom to blue.
Picovoice's output can be seen on the serial port monitor.
See below for the full context:
context:
expressions:
changeColor:
- "[turn, make] (all, the) lights $color:color"
- "[change, set, switch] (all, the) lights to $color:color"
- "[turn, make] (the) $location:location (color, light, lights) $color:color"
- "[change, set, switch] (the) $location:location (color, light, lights) to $color:color"
- "[turn, make] (the) [color, light, lights] [at, in] (the) $location:location $color:color"
- "[change, set, switch] (the) [color, light, lights] [at, in] (the) $location:location to $color:color"
- "[turn, make] (the) [color, light, lights] $color:color [at, in] (the) $location:location"
- "[change, set, switch] (the) [color, light, lights] to $color:color [at, in] (the) $location:location"
changeLightState:
- "[switch, turn] $state:state (all, the) lights"
- "[switch, turn] (all, the) lights $state:state"
- "[switch, turn] $state:state (the) $location:location (light, lights)"
- "[switch, turn] (the) $location:location [light, lights] $state:state"
- "[switch, turn] $state:state (the) [light, lights] [at, in] (the) $location:location"
- "[switch, turn] (the) [light, lights] [in, at] the $location:location $state:state"
changeLightStateOff:
- "shut off (all, the) lights"
- "shut (all, the) lights off"
- "shut off (the) $location:location (light, lights)"
- "shut (the) $location:location (light, lights) off"
- "shut off (the) [light, lights] [at, in] (the) $location:location"
- "shut (the) [light, lights] off [at, in] (the) $location:location"
- "shut (the) [light, lights] [at, in] (the) $location:location off"
slots:
color:
- "blue"
- "green"
- "orange"
- "pink"
- "purple"
- "red"
- "white"
- "yellow"
state:
- "off"
- "on"
location:
- "bathroom"
- "bedroom"
- "closet"
- "hallway"
- "kitchen"
- "living room"
- "pantry"
- Copy the UUID of the board printed at the beginning of the session to the Serial Wire Viewer (SWV).
- Go to Picovoice Console to create models for Porcupine wake word engine and Rhino Speech-to-Intent engine.
- Select
Arm Cortex-M
as the platform when training the model. - Select
STM32
as the board type and provide the UUID of the chipset on the board.
The model is now being trained. You will be able to download it within a few hours.
- Download your custom voice model(s) from Picovoice Console.
- Decompress the zip file. The model file is either
.ppn
for Porcupine wake word or.rhn
for Rhino Speech-to-Intent. - Use binary_to_c_array.py to convert your binary models to C array format utilizing the following command:
python3 binary_to_c_array.py input_binary_model output_c_array.txt
- Copy the content of
output_c_array.txt
and update thekeyword_array
andcontext_array
values in /stm32f407g-disc1/Inc/pv_params.h.