Collaborative Robots in Construction: A Proactivity Perspective

Abstract

Achieving effective human–robot collaboration (HRC) in dynamic environments like construction requires robots to function as proactive peers, anticipating human intents and acting in advance. While current literature analyzes enabling technologies through a sense–plan–act lens or catalog construction applications, the literature predominantly focuses on the robot’s internal autonomy pipeline. Studies lack a systematic examination of robot proactivity—specifically, how robots perceive, interpret, and anticipate human intent. The question of whether we are really making progress toward proactive human-centered collaboration remains inadequately addressed in existing HRC reviews. To answer this question, we propose a five-level taxonomy of robot proactivity: Reactive (L1: Explicit command-driven), Responsive (L2: Implicit command-driven), Predictive (L3: Behavioral intent-aware), Empathetic (L4: Human-aware), and Symbiotic (L5: Shared intention). This taxonomy is grounded in two key dimensions: autonomy (robot IQ) and intention anticipation (robot EQ), providing a nuanced understanding of interaction dynamics and collaboration maturity trends. Through a systematic analysis of 130 research papers from Scopus and Web of Science published over the past two decades, we reveal that while progress has been made in reactive, responsive, predictive, and empathetic capabilities, construction robotics still fall short of achieving symbiotic collaboration. By identifying gaps at each level of robot proactivity, this review proposes actionable research directions, such as vision-language-action (VLA) models, uncertainty-aware adjustable autonomy, and team contingency anticipation, paving the way toward more proactive collaboration. This review contributes to the field of HRC in construction by introducing a unified, human-centered taxonomy of robot proactivity grounded in autonomy and intention anticipation. By reframing HRC through the lens of proactivity rather than technical functionality alone, this work provides both a conceptual foundation for advancing intelligent collaborative systems and a practical roadmap for researchers and industry stakeholders aiming to realize truly proactive robotics on construction sites.