Routine Computing: A Systematic Review of Sensing Daily Life Dimensions Towards Human-Centered Goals

A team from Tsinghua University (Borislav Pavlov, Jiajin Li, Jun Fang, Yuntao Wang, Yuanchun Shi) published the first systematic review of routine computing, analyzing 203 studies up to August 2025. The paper introduces a new taxonomy focusing on temporal structures, behavioral interactions, cognitive aspects, and how variability and deviations are addressed. It identifies four major application domains: accessibility care (e.g., supporting elderly or disabled individuals), promoting healthy habits (e.g., fitness or sleep tracking), adaptive and context-aware support (e.g., smart home automation), and large-scale population insights (e.g., urban planning or public health).

Persistent challenges are highlighted, including the gap between low-level activity recognition (e.g., detecting a user is walking) and high-level intent (e.g., understanding they are heading to a meeting), the tension between personalization and generalization across diverse users, unresolved privacy concerns with continuous sensing, and data-related limitations like labeling costs and sensor noise. The paper provides a foundational framework for HCI researchers, outlining principles for designing ethical, adaptive, and human-centered routine-aware systems. It is set to be published at the ACM CHI 2026 conference.