Talking trash is a system for smart trash management. It enables trash cans to be equiped with multiple different sensors and reports this data to a server. The server aggregates and analyzes the data to provide useful statistics. This is then used to map optimal routes for emptying the trash cans effectively.

The trash cans uses two ultrasound sensors for measuring the distance from the lid of the trash can to the bottom. These are used to measure how full the trash can is. The reason for having two of them is in case the trash is unevenly distributed, so there still is some space on one side. It also uses a flame sensor in the lid to measure if the trash is taking fire. These sensors are connected to an Arduino Nano, which runs the code provided in `arduino/lid. The last sensor in the lid is a TI SensorTag. This has an accelerometer, which is used to detect if the lid is open or closed.

In the bottom of the trash can, we have some weight sensors. This is useful to get an idea of the types of trash in the can, and how much it can be compressed. If the can is full, but very light, we can assume that the contents can be compacted by the garbage truck, which means we can empty more trash cans in one round. This sensor will be connected to an Arduino Uno placed in the bottom of the can, which runs the code provided in arduino/weight.

The two Arduinos and the SensorTag are connected to a Raspberry PI placed on the back side of the trash can. The Arduinos are connected with USB and communicates over a serial bus. The SensorTag is connected with Bluetooth. The Raspberry PI runs the code provided in raspi to collect the data and push it to the server.

####Note: there are a few util scripts in the scripts folder, go chek them out to fill your system with data!

Prerequisites:

• Setup a DocumentDb in Azure
• Google Maps ApiKey

Replace Azure DocumentDB settings in appsettings.json:

{
  "ConnectionStrings": {
    "DefaultConnection": [LocalDBConnectionString]
  },
  "DatabaseSettings": {
    "Collection": [CollectionName],
    "Database": [DbName],
    "AuthKey": "[AuthKey]",
    "Endpoint": "[EndpointUrl]"
  }
}

Replace googleApiKey in Views/HeatMap/Index.cshtml:

<script src="https://maps.googleapis.com/maps/api/js?key=[GoogleApiKey]&libraries=visualization&callback=initMap" async defer></script>

Replace Azure DocumentDB settings in Web.config:

<appSettings>
  <add key="endpoint" value="[Endpoint URL]" />
  <add key="authKey" value="[AuthKey]" />
  <add key="database" value="trashTalk" />
  <add key="collection" value="[CollectionName]" />
</appSettings>

To run the Arduinos, first connect the sensors to the pins defined in the files in the arduino directory. We use an Arduino Nano for lid and an Arduino Uno for weight. Then use the Arduino SDK to open each of the .ino files and upload them to the Arduinos. The Arduinos will run automatically when given power, so after uploading you just need to connect them to the Raspberry Pi with USB.

The Raspberry Pi needs to be registered with the server. To do this, register a new device in the server and copy the device id to ~/.config/trashtalk/device_id on the Pi. Then find the MAC address of your TI SensorTag and set SENSORTAG_MAC in raspi/client.py to this address. You can set the API endpoint with API_PREFIX.

To run the program, you need to install some dependencies first. You can do this by running pip install -r requirements.txt. After setting the config and installing the dependencies, you simply need to run the program with ./raspi/client.py.

The Android App is set up as an Android Studio project. To run it you can import the project in Android Studio and install it on a device. You can change the API endpoint by editing API_PREFIX in MainActivity.java.