Group 3 : Cognitive Living

Group 3 : Cognitive Living

The Team

* Carl Kugblenu * Geraldine Villers * Madhubala Ganesan * Syed Asif Iqbal

Motivation

Increased use of Electronic devices for the household
Existing Home Automation systems are designed for specific purposes – comfort / security / cost saving
Need for an integrated system which serves all these purposes
Users’ needs change from time to time. So there is a need for self adapting system
Social – Comfort & Security; Environmental – CO2 emission; Economic – Cost saving

Goals

A comfortable environment – physiologically ambient environment for living
Safe and Secure living
Energy aware living & steps towards saving energy
Cost saving by reducing power consumption
A system that can automatically understand user’s needs

Our Idea

We have implemented a system with DHS HomeManager, FHEM and HomeMatic devices which can provide comfort, security and save energy in a household. We have three operating modes namely : Normal Mode, Shield Mode and Child Mode. Normal Mode is when people are living in the house, so the focus of this mode is comfort. Shield Mode is when no one is at home, so we focus on energy saving & security. Child mode is where a child is left at home alone. So the focus here is comfort, safety & security of the child. We control the AirConditioner / Heating system based on room temperature and humidity values. With the help of Axis M115 camera we show the camera feed to the user at all times. We also detect motion in specific areas of the frame and alert the user in case of abnormal activities in front of the door or if there is an attempt to tamper the camera by email. The system alerts the user if doors and windows are open in Shield & Child mode for safety & security. In addition, the system should be able to predict and adapt itself based on user's behaviour.

Components Used

DHS HomeManager
FHEM
HomeMatic Wireless Door / Window Sensor - optical
HomeMatic Wireless switch Actuator 1-channel
HomeMatic Wireless Temperatue / Humidity sensor indoor
AXIS M1113 Network Camera

System Architecture

Cognitive Prediction (Code Snippet)

This is the Predition service code sample to train the neural network model with collected home data

var brain = require('brain');
var fs = require('fs');

// Training data for temperature, humidity and switch response
var training_data = JSON.parse(fs.readFileSync('training.json', 'utf8'));
var net = new brain.NeuralNetwork();

// Training the neural network
net.train(training_data);

// Predict a switch response based on a temperature and humidity input
var output = net.run([0.22, 0.11]);     // 0.00112''

FHEM Definitions (Code Snippet)

Normal Mode

define NormalMode notify DoorWindowSensor {
   if (($EVENT eq "closed") && (ReadingsVal("TempHumiditySensor","temperature","") >= 25.0)) {
        fhem "set Switch on"
        } 
   else {
       fhem "set Switch off"
        }
  }

Shield Mode

define ShieldMode notify MotionSensor {
    if (($EVENT eq "motion")) {
        fhem "set Switch off"
        }
  }

Child Mode

define ChildMode notify DoorWindowSensor {
     if (($EVENT eq "open")) {
         fhem "set Switch on"
         }
     else {
         fhem "set Switch off"
         }
  }

It Works!

Home automation protocols

LonWorks

What is LonWorks?

Features of LonWorks

Efficient delivery of small messages.
Reliable delivery of messages.
Duplicate message detection.
Multiple communications media.
Low installation and maintenance cost
Efficient use of channel bandwidth.
Prevent tampering.

Applications

Municipal and highway/tunnel/street lighting.
Heating and air conditioning systems.
Subway train control.
Stadium lighting and speaker control.
Security systems, fire detection and suppression, and newborn location monitoring and alarming.
Smart electricity metering

DASH7 Alliance Protocol

Features

Is an open source Wireless Sensor and Actuator Network protocol.
Built on top of an asynchronous Wireless Sensor Network Media Access Control.
This standard was mainly used for military logistics.
It uses Generally Radio Frequency Identification technology, which is used for locating and tracking objects in a variety of applications.
This technology consists of a tag and a reader. There are two types of tags, active and passive. Passive tags receive their energy from the readers, when readers emit electromagnetic radiations to detect the tag.

BLAST Network Technology