Murshed J. Ahmed      (mja2196 at columbia.edu)
Shijun "Scott" Hou    (sh3658 at columbia.edu)
Robert Fea            (rf2638 at columbia.edu)

• This project uses a Raspberry Pi 3 Model B.
• The operating system is Raspbian, installed with the NOOBS installer.
• An HDMI Cable and HDMI capable monitor is recommended for initial setup and debugging.
• Configure a wireless network connection.

  • Edit the file /etc/wpa_supplicant/wpa_supplicant.conf

    country=US
    ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev
    update_config=1
    
    network={
        ssid="My Unsecured WiFi Network"
        key_mgmt=NONE
    }
    

• Generally they can be installed with sudo pip install package_name_here
• Adafruit packages require manual setup with instructions here.

Package Name      Version
----------------- -------
Adafruit-ADXL345  1.0.1
Adafruit-GPIO     1.0.3
Adafruit-PureIO   0.2.1
boto3             1.6.18
botocore          1.9.18
gpiozero          1.4.1
numpy             1.12.1
picamera          1.13
pip               9.0.1
requests          2.18.4
RPi.GPIO          0.6.3
simplejson        3.10.0
urllib3           1.22
weather-api       1.0.2

Can be found after cloning this repository

• Contains deployment pacakge for AWS Lambda function that generates nutrition reports.
• Files in this directory should be imported into AWS Lambda.

• Contains throwaway or one-time use only scripts
  • Testing API Calls
  • Hardware Interfacing Examples
  • Experimenting with component configurations and settings

• Contains proper project source code for running the main program.
  • src/sensors contains modules that abstract functionality for sensor inputs
  • src/output contains modules that abstract functionality for outputs
  • src/api contains modules that interact with third party APIs
  • config.txt should be updated with AWS keys

• Raspberry Pi Board
• Accelerometer
• Camera
• Trash Can
• Breadboard
• Dupont Breadboard Wires
• Board Mounting Plate
• Servo Motors (two)

• Edit the /etc/rc.local file to run commands at start up.
• You may wish to include one for emailing Raspberry Pi's IP address at boot time
• Assuming you have cloned the repository to /home/pi/Documents/ican-project
  • Add the following line to /etc/rc.local to begin executing the main program on boot and dump the output to a file:
  • /usr/bin/python /home/pi/Documents/ican-project/src/ican.py &> /tmp/ican.txt

• Refer to the src/sensors/lid.py module.

• Connect the following pins from the Raspberry Pi to the ADXL345 Accelerometer

  Raspberry Pi      ADXL345
  3.3 V             3.3 V
  GND               GND
  SCL1 (Pin 5)      SCL
  SDA1 (Pin 3)      SDA
  3.3 V             CS
  GND               SDO
  -                 INT1
  -                 INT2
  

• Look for an LED lighting up when connected.

• Using the ribbon connector tape connector, insert one end to the Pi Camera
• Insert the other end to the Raspberry Pi's Camera port
  • This is between the HDMI connector and the audio jack
  • Labelled on the PCB as CAMERA

• This is simply and LED to indicate if the Lid is open or not
• Positive lead goes to Raspberry Pi GPIO 21 (Pin 40)
• Negative lead is connected to a small resistor (approx 100 ohms)
• Resistor is connected in series to common GND

• Servos are mounted and screwed down on the back of the trash can, using the mounting flanges.
• A wide hole must be drilled for the output axis into the trash can.
• Two flaps made of corrugated plastic sheets must be cut for the trapdoor.
• The GND wires of both servos must go to common ground (of the Raspberry Pi)
• The VCC wires of both servos must go to an external battery source providing approximately 7 volts
• The signal wires of each servo must go a GPIO pin on the Raspberry Pi.
  • Refer to src/output/trapdoor.py.
  • GPIO 18 (Pin 12) and GPIO 24 (Pin 18)

• Mount the Raspberry Pi, Breadboard, Accelerometer and Camera to Board Holder.
• Screw the Board Holder into the lid of the trash can.
• Insert USB Power Cable through the back of the trash can (or create your own hole).
• Once powered on, the board should begin running the main program.
  • You can tell the program is running by the Camera LED being on

• Sign up for a developer account with Nutritionix.
• Replace your API keys in lambda/nutrition.py

• Sign up for an AWS Account. Free Tier is sufficient.
• Create a Role with Cognito and IAM
• Replace the COGNITO_ID in src/api/aws.py
• In the file src/config.txt place your keys from your AWS account.
  • The values should be one per line
    • Account ID
    • Identity Pool ID
    • Role ARN

• Simply create a new bucket for the project and update the information in src/api/s3.py

• Create a new classification model in AWS ML
• Use the training data format in test_scripts/training_data_ican.csv
• Create a realtime endpoint and update the file in src/api/prediction.py

• Create a new database table
• Update the name in src/api/database.py

• Create a new topic in SNS.
• Add yourself (your cell phone number) as a subscriber.
• Update the information in src/api/notification.py

• Create a new Lambda function making sure that it has a role with full access to
  • DynamoDB
  • SNS
  • S3
• Import the deployment package files in the lambda directory

• Run the command feh /path/to/my/image.jpg
• Can also use wildcards to view multiple images
  • feh pic*
    • Would match pic, picture, pic_001, etc.
  • Use left/right arrow keys to navigate
  • Images are generally loaded in alphabetical order

• Under normal circumstances, with a stable network, the iCan should immediately start running once plugged in
  • You can tell the program is running if the red camera light is on and the "Open Indicator" LED responds to movement
• If the program is not running but you know the IP address
  • SSH into the Raspberry Pi
  • Clear the /tmp directory of all text files
  • Run the main iCan program python /home/pi/Documents/ican-project/src/ican.py
• If the program is not running and you don't know the IP address
  • Connect the portable monitor via HDMI
  • Open a terminal and run the same commands above
  • Alternatively you can get the IP address of the wlan0 interface (use the command ifconfig)
    • Once you have the IP address you can SSH with the credentials above