AI-Driven Activity of Daily Living Monitoring: A Comprehensive Review of Ambient, Wearable, and Fusion-Based Sensing Technologies.

The increasing prevalence of chronic diseases, functional decline, and cognitive impairments, together with the limitations of episodic clinical assessments, has created a critical need for automated and continuous monitoring of Activities of Daily Living (ADLs) to support timely intervention, personalized care, and independent living. This review systematically examines how ambient sensing, wearable devices, and fusion-based systems have been applied to ADL monitoring in home environments. We analyze 109 peer-reviewed studies published between 2013 and 2024, focusing on reported performance trends, study scale, sensing configurations, and methodological characteristics. Given the substantial heterogeneity across experimental settings, participant populations, activity definitions, and evaluation protocols, reported metrics are interpreted descriptively rather than as directly comparable measures. Overall, wearable-based systems frequently report higher performance metrics in controlled settings, ambient systems offer advantages related to privacy and unobtrusiveness, and fusion-based approaches provide richer contextual information but face scalability challenges. Across all modalities, limited dataset diversity, small sample sizes, inconsistent evaluation practices, and insufficient reporting of real-world deployments remain persistent gaps. This review synthesizes recent advances in sensor-based activity recognition for monitoring activities of daily living, with a particular emphasis on AI-driven approaches.