Human Activity Recognition using Inertial, Physiological and Environmental Sensors: A Comprehensive Survey

Area of Science:

• Computer Science
• Biomedical Engineering
• Artificial Intelligence

Background:

• Human Activity Recognition (HAR) has gained prominence due to widespread electronic devices like smartphones and smartwatches.
• Advancements in machine learning, particularly deep learning, have fueled HAR applications across diverse fields such as sports and healthcare.
• HAR is a key assistive technology for monitoring elderly individuals' cognitive and physical functions through daily activities.

Purpose of the Study:

• This survey critically examines the role of machine learning in developing HAR applications.
• The focus is on HAR systems utilizing inertial sensors combined with physiological and environmental sensors.

Main Methods:

• The study surveys existing literature on machine learning algorithms for HAR.
• It specifically analyzes approaches that integrate data from inertial sensors (e.g., accelerometers, gyroscopes) with physiological sensors (e.g., heart rate monitors) and environmental sensors (e.g., temperature, light).

Main Results:

• Machine learning algorithms are crucial for interpreting complex sensor data to accurately recognize human activities.
• Sensor fusion, combining data from multiple sensor types, enhances HAR system performance and robustness.
• The integration of diverse sensors provides a more comprehensive understanding of user behavior and context.

Conclusions:

• Machine learning is fundamental to the advancement of effective Human Activity Recognition systems.
• The combination of inertial, physiological, and environmental sensors offers significant potential for sophisticated HAR applications.
• HAR technology holds substantial promise for improving health, well-being, and assistive care, especially for the elderly.