ARIL (Activity Recognition Indoor Localization) is a public benchmark dataset designed to support CSI-based Wi-Fi sensing research, providing labeled Channel State Information samples for the joint or independent tasks of indoor human activity recognition and localization. It matters to the field because it offers a standardized, reproducible resource that enables fair comparison across models and contributes to the broader effort of making Wi-Fi sensing research more transparent and verifiable. While the dataset primarily targets single-user scenarios, its dual-task structure — covering both activity classification and positional inference — distinguishes it from datasets focused on only one of these sensing objectives.
Dictionary term