Muhammad Raisul Alam

A Review of Smart Homes—Past, Present, and Future

A smart home is an application of ubiquitous computing in which the home environment is monitored by ambient intelligence to provide context-aware services and facilitate remote home control. This paper presents an overview of previous smart home research as well as the associated technologies. A brief discussion on the building blocks of smart homes and their interrelationships is presented. It describes collective information about sensors, multimedia devices, communication protocols, and systems, which are widely used in smart home implementation. Special algorithms from different fields and their significance are explained according to their scope of use in smart homes. This paper also presents a concrete guideline for future researchers to follow in developing a practical and sustainable smart home.

SPEED: An Inhabitant Activity Prediction Algorithm for Smart Homes

This paper proposes an algorithm, called sequence prediction via enhanced episode discovery (SPEED), to predict inhabitant activity in smart homes. SPEED is a variant of the sequence prediction algorithm. It works with the episodes of smart home events that have been extracted based on the on -off states of home appliances. An episode is a set of sequential user activities that periodically occur in smart homes. The extracted episodes are processed and arranged in a finite-order Markov model. A method based on prediction by partial matching (PPM) algorithm is applied to predict the next activity from the previous history. The result shows that SPEED achieves an 88.3% prediction accuracy, which is better than LeZi Update, Active LeZi, IPAM, and C4.5.

Human activity classification for smart home: A multiagent approach

Smart home research requires study of psychological characteristics of home user. People follow some specific patterns in their life style. Inhabitant activity classification plays a vital role to predict smart home events. The paper proposed a multiagent system to track the user for task isolation. The system is composed of cooperative agents which works by sharing local views of individual agents. An algorithm is derived based on opposite entity state extraction for activity classification. The algorithm clusters the smart home events by isolating opposite status of home appliance. Result shows that the proposed algorithm can successfully identify inhabitant activities of various lengths.

A spatiotemporal model of human circadian rhythm in smart homes

This article presents a spatiotemporal model of human circadian activity rhythm in smart homes. A spatiotemporal model is used to represent human activity in a time-based system. This article proposes a learning and prediction algorithm to analyze temporal characteristics of the residents activity. The algorithms combined Allens temporal logic and Gaussian distribution to incrementally learn and predict next activity of the inhabitant. The methods show 88.1% prediction accuracy when tested with a practical smart home data set. Further analysis showed that human activity in smart homes follows Gaussian distribution, which previously had been merely an assumption.

A Task Oriented Multiagent Model for Smart Home Utilizing Hierarchical Agent Architecture

Smart home is an extension of modern home automation system which possesses computational intelligence to solve problems in ubiquitous environment. Multiagent algorithm is frequently used to reduce uncertainty in pervasive environment. This research involves designing a multiagent model which consists of distributed task organizing agents. Each agent is responsible for specific task which is partially shared with a supervisor agent for final prediction. The agents are organized for time, place and event prediction. Agent architecture follows a layered approach with interconnected processing units. The system provides an efficient method to predict events with temporal characteristics based on user location.