Over the last few decades, smart home systems have failed to spread widely in every home and be a coherent part of our life like what smart phones have done in half this period of time. On the other hand, the latest evolution of Internet of Things (IoT) and Big Data analysis gave new insights of smart platforms that can potentially lead to the new dream of ‘smart cities’.
It is expected that the most used daily tools and objects (Things) like clothes, brushes, keys, doors, etc. will be equipped with smart chips to be transformed into IoTs for future activities. Therefore, it makes perfect sense to think of future smart homes as huge virtual networks of Things that collect and process massive amount of sensor data and not just a few number of sensors like those introduced as individual products from IBM, Samsung and Intel. In our research, we survey the state-of-the-art of basic components of smart home systems, and introduce a framework for developing adaptive and self-learning home automation systems using both IoT and Big Data analytics.
The proposed framework is to augment homes with machine-learning intelligence to automatically detect different situations, self-learn how inhabitants act and eventually react and control the home autonomously. We propose our SLASH (Self-Learning and Adaptive Smart Home) framework for designing and implementing smart home systems that are both adaptive and self-learning. Our framework suggests integrating IoT across every home with a large network connected to Big Data analyser. Such an engine that supports analysing inhabitants’ behaviours on a large-scale can enable a new type of home automation that depends on machine learning and develops on-going automation decision over time.
For example, the user turns on the entrance light when entering his home – without any change in the status of other things- then after a defined number of reoccurrences, the system will sense the user’s entrance and automatically turn on the lights. The situation is defined per each sensor status – within the home –and according to the change(s) that led to that situation. With the increase in number of situations and the complications, SLASH can learn and act to support the inhabitant’s routine life.
SLASH is different than all previous work in smart home field in the sense that it starts without any predefined configuration, but rather as a new born baby gaining experiences from different situations as it grows. SLASH framework utilizes the relatively consistent behaviour of different users, captures and stores IoT/Things sensor data, and analyses through machine learning and big data analytics how the smart home should interact with inhabitants.
The framework is described in more details in an IEEE published paper stating the different stages of data collection, manipulation and data correlation that could potentially provide people with a better experience and quality of living.
If you’re interested in learning more about this or would like to indulge in some conversation, do drop us a line!