Skip to content
/ iot-project-2020 Public

IoT platform for user-involved air recirculation system.

3 stars 1 fork Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

MatteoBattilana/iot-project-2020

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

246 Commits
services
services
 
 
.gitignore
.gitignore
 
 
README.md
README.md
 
 
diagram
diagram
 
 
proposal template v0.2.pdf
proposal template v0.2.pdf
 
 

Repository files navigation

CleanAir

Politecnico di Torino - Final Project for the course 'Programming for IoT', AY 2020/2021

Description of the infrastructure CleanAir

Promo video: CleanAir Promo

Demo video: CleanAir Demo

The entire infrastructure is based on a central service that is called Broker. It keeps a list of available services with their corresponding REST and MQTT endopoints in order to allows the other services to know which services are available. Within a time, that is called retationTime, each service must perform a ping operation that is nothing else than a notification to the catalog to say "I'm still alive". In this way, the record for that service will not be removed from the catalog. This is done by saving an additional information in the json of each service called lastUpdate that contains the time in which the last ping has been performed.

Eache service is indentified by and id (SIMULATED-DEVICE-1) but it is defined also by two attributes:

  • TYPE: it can assume only two values: [SERVICE, DEVICE]
  • SUBTYPE: ideally it can be every type, depends on the alive services in the infrastructure, at the moment we have [EXTERNALWEATHERAPI, WEBINTERFACE, ...]
  • groupId: only the services with TYPE = DEVICE have this attribute. Since the infrastructure must be scalable and maybe in the future, this will be used by more than one user, the groupId attribute define to which room the device belong. For example, if we have two users: Marco and Matteo, the sensor that are inside Marco's house will have a different groupId. At the moment this has been implemented in the code and works correctly but no simulated sensors have been implemented. This is because this is not actually requested.

Update: at the startup the sensor is not associated to any groupId, via Telegram the user must configure a new device by calling the /setGroupId API of the sensor by passing the new groupId and the pin, that works like a password. The id of the service/device is set via a json configuration file.

This is an extract from a http://localhost:8080/catalog/getAll response request:

{
    "serviceName": "SIMULATED-DEVICE",
    "serviceType": "DEVICE",
    "serviceSubType": "RASPBERRY",
    "groupId": "home1",
    "devicePosition": "internal",
    "serviceServiceList": [
      {
        "serviceType": "MQTT",
        "endPoint": [
          {
            "topic": "/iot-programming-2343/",
            "type": "temperature"
          }
        ]
      },
      {
        "serviceType": "REST",
        "serviceIP": "172.20.0.4",
        "servicePort": 8080,
        "endPoint": [
          {
            "type": "web",
            "uri": "/",
            "version": 1,
            "parameter": [
              
            ]
          },
          {
            "type": "configuration",
            "uri": "/setPingTime",
            "version": 1,
            "parameter": [
              {
                "name": "pingTime",
                "unit": "integer"
              }
            ]
          },
          {
            "type": "configuration",
            "uri": "/setGroupId",
            "version": 1,
            "parameter": [
              {
                "name": "groupId",
                "unit": "string"
              }
            ]
          },
          {
            "type": "action",
            "uri": "/forceSensorSampling",
            "version": 1,
            "parameter": [
              
            ]
          }
        ]
      }
    ],
    "serviceId": "SIMULATED-DEVICE-2",
    "lastUpdate": 1627895704.1240659
  }

It's important to say that each service cannot directly communication each others since the ip of the service can change, but they must ask the catalog the ip/port and enpoints of a specific one. For example, if the sensor needs to perform a query about the externa weather condition, it must request to the catolog which is the ip and port for that desired service. So,

  1. http://catalog:8080/catalog/searchByServiceSubType?serviceSubType=EXTERNALWEATHERAPI
  2. Check if the list is not empty
  3. Take the last one and perform the desire request

At this point, one the catalog has been loaded, the new services will perform a ping to it in order to be registered in the catalog list of the active services. This is both for SERVICE and DEVICE. In order to simplify this, a class shared among all the services has been created and named ping.py.

Services description

More specifically the infrastructure is made up of 7 different services: the most important, on which all the system relies, is the Catalog; then the Device Adaptor, the External Api Adaptor, the Telegram Manager, the ThingSpeak Adaptor, NodeRed and finally the Control Strategy.

Catalog

It performs the dual function of Service Catalog and Device Catalog and it is the starting point for every service inside the infrastructure: in fact its purpose is to list all the services and devices endpoints and the resources they expose in order to be able to comunicate with them. Every service must perform regularly a ping operation to inform the catalog that it is up and running: in fact the catalog checks periodically if time from the last ping operation has expired and, if so, it removes the service from the list.

Telegram Manager

The Telegram Manager exposes RESTFUL APIs which are going to be used by other services: in particular the Control Strategy used in order to notificate the end-user.

Nodered

The Telegram Manager exposes RESTFUL APIs which are going to be used by other services: in particular the Control Strategy used in order to notificate the end-user.

Control strategy

It’s the infrastructure core service:

  • It receives and analyses data coming from the devices
  • It does not merely check that the data for the various measured quantities are above the threshold limit, but it performs a multiple control in order to avoid spurious data problems.
  • It alerts the user even if the devices measurements have not reached a critical value yet, but are going to do so: implementing a function which predicts the future measurements’ value based on the past ones, this service is able to understand if some kind of measure is going to be critical in the next future.
  • Through Telegram Manager REST end-points, it sends to the user the notification in case one of the measured quantities is critical. At this point, analysing whether the user has or not an external device linked to the platform, and especially based on the actual external conditions plus the air pollution levels, it suggests to the user what is the best action to do in order to restore an healthy environment.
  • If the external conditions are good and so is the air quality outside the user is told to open its window. Otherwise, in case one of these two conditions is not fulfilled, the control strategy tells the user not to open the window, but to open an internal door instead or eventually to switch on the dehumidifier; furthermore, based on the forecasted value provided by the external weather API, this service is able to inform the end-user if the external weather and the air pollution levels are going to return within normal ranges and when so that the user knows in advance when to open the window.

ThingSpeak adaptor

ThingSpeak is used in order to store, visualize, and analyze live data streams in the cloud. It provides instant visualizations of data posted by our devices. The Thingspeak Adaptor is our link between Thingspeak itsself and the infrastructure: it is subscribed to sensors topic in order to be able to receive their live messages and then it upload them to ThingSpeak via REST. It also performs daily, weekly and monthly statistics for all the measured quantities.

ExternalWeatherApi

The ExternalWeatherApi service contact OpenWeather API in order to get:

  • Current weather data
  • Air pollution data
  • Forecasted weather data
  • Forecasted air pollution values
  • These informations are exposed via RESTful APIs to other services and then processed by the Control Strategy.

Setup

Windows 10 setup

  • Download git: Git install tutorial
  • Download and install docker: Docker installer
  • Download and install WSL: Stack Overflow guide
  • Set the environment variable: Env variable tutorial. The variable names you have to set are (the key values must asked to @MatteoBattilana):
    • OPENWETHERMAPAPIKEY
    • THINGSPEAKAPIKEY
  • Clone the project repository to your computer:
    • Open the git bash just installed
    • Write the following command: git clone https://github.com/MatteoBattilana/iot-project-2020.git
  • In order to start the architecture you have to:
    • Enter the service directory: cd iot-project-2020/services
    • Build the architecture (it has to be done once unless you made some changes to Dockerfiles or requirements.txt files): docker-compose build
    • Start the architecture (it will take a while): docker-compose up

Fix node-red problem

You should enter the project directory, open the git bash terminal and execute:

git config --global core.eol lf
git config --global core.autocrlf input

GIT commands

Update the repo to the last update without modification to the same file

git fetch
git pull

Create a branch and enter

git branch newbranch
git checkout newbranch

Clean/Restore a local branch (discard your local modifications inside a someone else branch)

git reset --hard

Local save of your work in a branch (to do before switching to another branch or having done a commit)

git stash (push) --> save the modifications
git stash pop --> to have the modifications back

Push your changes to GitHub, to your branch

git add .
git commit -m "Added servicetest"
git push origin newbranch

Once you have done your work, open a pull request using the GitHub interface

Material links which can be useful

PROJECT REST ENDPOINTS

About

IoT platform for user-involved air recirculation system.

Topics

python docker mqtt iot arduino smarthome docker-compose

Resources

Readme
Activity

Stars

3 stars

Watchers

2 watching

Forks

1 fork
Report repository

Contributors 3

  •  
  •  
  •  

Languages