Sensor-Based Benchmark Datasets: Comparison and Analysis

Human Activity Recognition (HAR) using installed sensors has made renowned progress in the field of pattern recognition and human-computer interaction. To make efficient machine learning models, researchers need publicly available benchmark datasets. In this chapter, we have bestowed a comprehensive survey on sensor-based benchmark datasets. We have not considered RGB or RGB-Depth video-based action or activity-related datasets in this book. We have performed a complete analysis of benchmark datasets, that incorporates information about sensors, attributes, activity classes, etc. These datasets sum up a good number of sensor-based daily activities, medical activities, fitness activities, device usage, fall detection, transportation activity, and hand gesture data.