Towards a Bathroom-Centered Human–Building Digital Twin Framework for Indoor Safety Analysis

Older adults face significant safety risks in bathrooms due to the combination of wet surfaces, constrained layouts, and the physical demands of frequent posture transitions. Traditional approaches to this problem have been fragmented: built-environment studies focus on static design modifications (e.g., grab bars), while human-centered sensing studies focus on detecting falls without understanding the environmental context. This paper argues that bathroom safety is a dynamic, coupled process where human movement and environmental features interact, necessitating a more integrated analytical approach.

The authors introduce a six-layer digital twin (DT) architecture designed to bridge the gap between human motion and environmental context. The framework includes:

• Physical Sensing Layer: Captures real-time data from the bathroom environment.
• Human-Building Twin Construction: Creates a semantic model of the bathroom (fixtures, zones) and a skeleton-based representation of the occupant.
• Spatial-Semantic Coupling: Aligns the human skeleton with the bathroom coordinate system to map movements to specific objects or functional zones.
• Interaction-Aware Analytics: Extracts kinematic descriptors and contact events to interpret behavior in context.
• Visualization and Decision Support: Provides an interface for safety-oriented interpretation of these interactions.

By moving beyond simple fall detection, this framework allows researchers and designers to understand why a safety event occurs. For example, it can distinguish between a fall caused by a slippery floor versus an unstable transfer from a toilet. This shift from person-centered to person-in-environment-centered analysis provides a methodological foundation for developing privacy-sensitive, interaction-aware applications that support aging-in-place.

Digital Twin (DT)

A data-connected virtual representation that integrates physical geometry, real-time sensing, and analytical models to simulate and monitor real-world processes.

Skeleton-based Human Representation

A method of tracking human posture and movement using joint coordinates, which provides kinematic data while preserving privacy by avoiding raw video footage.

Spatial-Semantic Coupling

The process of mapping dynamic human motion data onto a semantically labeled 3D environment, allowing the system to understand which objects or zones are involved in a specific movement.

Environmental Press

The degree of difficulty or demand that a physical environment imposes on an individual, which varies based on the person's specific mobility and balance capacity.

AI-generated third-party summary by ResearchPod. Not official content or an endorsement by the paper authors or affiliated organizations.