Human Motion Capture for Physical Human-Robot Interaction: A Systematic Review

Abstract

This paper systematically reviews state-of-the-art motion capture systems in the context of physical Human-Robot Interaction (PHRI), where humans and robots are in direct physical contact and motion sensing directly affects control stability, interaction safety, and task performance. As robots are increasingly deployed in industrial and healthcare settings that require close and continuous human contact, reliable real-time estimation of human motion has become a critical bottleneck for safe and effective PHRI. However, existing motion capture solutions for PHRI are reported across disparate research communities—including robotics, biomechanics, computer vision, and wearable sensing—using inconsistent assumptions, evaluation metrics, and experimental contexts, making it difficult to compare methods or identify general design principles. This fragmentation motivates the need for a systematic review. Accordingly, 54 studies were analysed, as selected from an initial pool of 4,406 manuscripts, that focus on implementations and evaluations of human motion capture for kinematic analysis in PHRI scenarios. The findings show that current approaches predominantly rely on wearable or vision-based systems, despite persistent limitations related to occlusion, environmental sensitivity, calibration burden, and cost. While some studies exploit the robot itself as a high-fidelity sensing platform, the integration of robotic sensing into multi-modal sensor fusion frameworks remains underdeveloped, despite being well established in other areas of robotics. In addition, the lack of standardised evaluation protocols limits meaningful cross-study comparison. Overall, this review highlights the need for motion capture solutions that integrate robotic and complementary sensing modalities through robust sensor fusion, alongside standardised evaluation methodologies, to enable reliable and scalable motion estimation in real-world PHRI applications.