A Novel Human Activity Recognition and Prediction in Smart Home Based on Interaction.

Area of Science:

• Computer Science
• Artificial Intelligence
• Ubiquitous Computing

Background:

• Smart Homes offer solutions for elderly/disabled care and power saving.
• Human Activity Recognition (HAR) is crucial for smart home automation.
• Existing HAR methods lack scalability and struggle with non-default activities.

Purpose of the Study:

• To develop a novel framework for both recognizing and predicting human activities in smart homes.
• To address the limitations of existing HAR systems, including scalability and recognition of diverse activities.
• To propose a cost-effective solution for widespread smart home adoption.

Main Methods:

• A three-stage framework: recognition after activity, recognition during activity, and advance activity prediction.
• Utilizes passive Radio-Frequency Identification (RFID) tags for low-cost hardware implementation.
• Focuses on high scalability for real-world, non-laboratory applications.

Main Results:

• The proposed framework demonstrates effective performance in both activity recognition and prediction.
• Achieved high scalability, enabling practical application beyond controlled environments.
• The use of passive RFID tags significantly reduces hardware costs.

Conclusions:

• The novel framework enhances smart home capabilities by accurately recognizing and predicting human activities.
• Its scalability and low cost make it a viable solution for widespread smart home automation.
• This research paves the way for more intelligent and responsive smart home services.