Jim Sutherland developed the first home automation system as far back as 1966, which could control temperature, turn appliances on and off, and manage a shopping list.
Around 1984 the term “smart house” was introduced to describe the automation of heating, security and garden watering systems, as well as doors and lights.
Currently, home automation is widespread. Many people control their TV, alarms, lights, garage and house doors remotely with their cellphone. Home appliances such as the fridge and cooking devices have wi-fi modules enabling them to communicate.
Smart houses are connected to the internet and can, therefore, be monitored from a remote location. Effectors can be used to automatically turn the lights and appliances on/off, water plants, feed pets, and close the curtains or blinds at certain pre-set times or conditions.
However, simply connecting appliances or devices to the internet so it can be controlled by a smartphone does not create a smart home. The internet connection is essentially nothing more than a long-range remote control. A smart device needs to be intelligent and should be able to learn and/or take independent action.
Artificial intelligence (AI), and specifically machine learning and predictive modelling are thus becoming ever more popular in intelligent home systems, making home automation applications adaptable to their environments.
Several of the well-known, voice-activated systems now contain virtual assistants with the ability to learn the preferences and behaviour patterns of the residents in order to personalise the system.
Like a competent assistant, AI knows which member of the household is speaking, adjusts to each person’s personal preferences and, based on their behavioural patterns, can predict what they want or need even before they ask for it.
AI automates tasks without the residents telling it what to do.
Intelligent houses are progressing from basic roles like completing tasks to more proactive roles like preventing break-ins, checking your to-do list and making appointments - all tasks that AI can handle independently. To some extent, an intelligent home system is like an experienced butler. You do not need to teach the butler the basics of how a household works. The same is true for an AI home solution that can analyse data, learn from it and make decisions.
Picture the following scenario: Last night you forgot to set the alarm, but the home system knows your schedule and the travelling time to the office based on internet traffic reports for Fridays and set the alarm accordingly. When you get out of bed, the coffee is just finished brewing.
In your hurry, you don't lock the front door, but it's no problem as the home system knows you have left and locks the door; turns off the coffee maker, TV and lights; and sets the alarm system. While you are at work, the multiple sensors in your cupboards count the supplies, while the fridge checks the expiry dates and quantities of certain products. Since you are running low on some things, the home system places an online order.
When the order is delivered, the cameras at the front door recognise the delivery truck and open the front door after taking a picture of the deliveryman. A voice over the speakers asks him to put the packages inside the house. Cameras will continuously monitor the deliveryman until the door closes behind him. The home robot carries the packages to the kitchen and unpacks the items, while the automatic vacuum cleaner vacuums the house.
During the day the smart outdoor security cameras recognise a stranger walking up to the front door and immediately ensure that everything is locked and the alarm is set. If the person attempts to break in, the security company and you are notified immediately.
Just before you arrive from work, the home system turns on the air conditioning and your favourite music. You ask the AI system to show recipes based on the food you have in the house. The system presents a variety of recipes based on your food preferences. While you are having dinner, a voice announces over the speakers that the big game is starting in 15 minutes and pre-tunes the TV for you.
Like the refrigerator, the TV can display who is at the front door or bring up the picture from the security cameras or baby monitor.
Based on behavioural patterns, the dishwasher starts automatically when it senses that it is packed, selecting the programme according to the dirtiness of the content.
It may sound like science fiction, but all these technologies already exist. Several manufacturers have committed to AI and developed washers that learn how you like to wash certain types of clothing. The washer and dryer also communicate via the internet to make adjustments to the wash and dry cycles based on changes in the weather as predicted online.
MIT researchers developed an automated system, Duet, which track occupants within tens of centimetres by measuring the reflections of wireless signals off their bodies and pinging nearby mobile devices to predict their identity. This allows air conditioners to adjust the temperature based on the preference of the person in the room.
Since the home system knows who is moving around in the house and can compare it to an established baseline of behavioural data, it can easily detect when the activities are abnormal and notify a caregiver.
The system, for instance, can detect that a person is moving around more or less than usual and may be suffering from depression or another ailment. Such information, when collected over time, could be valuable for monitoring and treatment.
Artificial Intelligence is a crucial element that makes homes “intelligent” so that they can simplify life and eliminate clutter by automatically adapting to our needs.
Professor Louis Fourie is the deputy vice-chancellor: Knowledge & Information Technology at the Cape Peninsula University of Technology.
The views expressed here are not necessarily those of Independent Media.