treitmaier / SpeechBox

SpeechBox

SpeechBox is an information appliance to engage communities, and collect spoken-language responses, on topics of interest.

The system's hardware design and software are all released under the Apache 2.0 license. The hardware enclosure is adapted from the StreetWise Toolkit.

The SpeechBox Appliance is implemented as a State Machine and can easily be extended as well as customised with bespoke Audio Prompts. The Appliance interacts with the SpeechBox Server.

Component List

• Raspberry Pi 3B+
• Power supply
• SD card
• AIY Voice Kit V1
• Mini OLED display
• 12-key keypad

Wiring

Follow the instructions on AIY Voice Kit on how to wire up the button (with integrated LED), microphone, and speaker.

Keypad

Next connect the keypad.

Keypad Pin	BCM PIN	Physical PIN	VoiceHAT
Row 1	17	11	Driver 1
Row 2	27	13	Driver 2
Row 3	22	15	Driver 3
Row 4	26	37	Servo 0
Column 1	13	33	Servo 2
Column 2	12	32	Servo 4
Column 3	24	18	Servo 5

Then configure the Raspberry Pi to recognize the keypad as a keyboard. Use the device tree overlay in the repository. If you change the wiring, make sure you adapt the device tree overlay accordingly.

OLED Display

And finally wire up the OLED display.

Display Pin	BCM PIN	Physical PIN	VoiceHAT
Ground		6	GND (SPI)
Voltage In		1	3.3V (SPI)
3v3	Nothing	Nothing	Nothing
Chip Select	8	24	CE0
Reset	2	3	SDA (I2C)
Data/Command	3	5	SCL (I2C)
Clock	11	23	CLK (SPI)
Data	10	19	MOSI (SPI)

Enclosure

The hardware is encased in a laser-cut box made of 3mm acrylic.
The colours of the laser-cut box template correspond to different modes and orders of cuts.

Order	Colour	Mode	Speed	Power
1.	Red	Engrave	200	100
2.	Blue	Cut	14	80
3.	Black	Cut	14	80

OS Setup

Create The SD Card

1. Download & Flash AIY Voicekit Image or Raspbian Stretch
2. Enable ssh & wifi & keypad

• Create an empty file, named ssh, in the boot partition of the microSD.
• If you're connecting to the Pi using wifi, create another file name wpa_supplicant.conf in the boot partition:

  country=US
  ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev
  update_config=1
  
  network={
      ssid="your_real_wifi_ssid"
      scan_ssid=1
      psk="your_real_password"
      key_mgmt=WPA-PSK
  }
  

3. Configure keypad device tree overlay driver

Copy the 3x4matrix.dtbo file to the =/boot/overlay= directory of boot partition¹.

4. Edit config.txt file in boot partition to include:

# Enable 3x4 matrix keypad
dtoverlay=3x4matrix

# Enable SPI for display
dtparam=spi=on

Boot up and connect to the pi using ssh

Follow the instructions on AIY Voice Kit if you're unsure how to do this.

1. make sure you change the password passwd

2. Install all dependencies

sudo apt install oracle-java8-jdk sudo apt install input-utils sudo apt install git sudo apt install wiringpi

3. Test keypad

lsinput

/dev/input/event0
   bustype : BUS_HOST
   vendor  : 0x0
   product : 0x0
   version : 0
   name    : "MATRIX3x4"
   bits ev : EV_SYN EV_KEY EV_MSC EV_REP

Next check if the button presses set off key pressed and released events. input-events 0

/dev/input/event0
   bustype : BUS_HOST
   vendor  : 0x0
   product : 0x0
   version : 0
   name    : "MATRIX3x4"
   bits ev : EV_SYN EV_KEY EV_MSC EV_REP

waiting for events
14:21:51.857979: EV_MSC MSC_SCAN 0
14:21:51.857979: EV_KEY KEY_N (0x31) pressed
14:21:51.857979: EV_SYN code=0 value=0
14:21:51.947982: EV_MSC MSC_SCAN 0
14:21:51.947982: EV_KEY KEY_N (0x31) released
14:21:51.947982: EV_SYN code=0 value=0

Compile / Deploy software

The build.gradle.kts tiggers a 'fatJar' to be created as part of the gradle build process: build/libs/SpeechBox-1.0-standalone.jar

You'll need to copy this file to the home directory of the default user (pi) on the Raspberry Pi. The build_and_deploy-sample.sh illustrates how to automate this step.

The speechbox.service systemd file, which should be copied to the /etc/systemd/system/speechbox.service path, is responsible for launching the SpeechBox software.

After installing the speechbox.service systemd file, you can interact with the service using the following commands.

# To start the speechbox service
sudo systemctl start speechbox

# To check the status of the speechbox service
sudo systemctl status speechbox

# To enable the speechbox service (run at boot)
sudo systemctl enable speechbox

The file also contains environment variables that you'll need to customise:

• API_PASSWORD=password
• API_URL=https://speechbox.example.com

If you are developing the software locally, you should customise both environment variables as appropriate and consider setting the MOCK environment variable to a non-null value so that you don't need to deploy to the Raspberry Pi.

License

Apache 2.0

Footnotes

1. Or compile the 3x4matrix.dts file if you change the wiring from what is specified in the Keypad subsection above. ↩