Network construction strategies for the activation of autonomous mobile robots

Autonomous mobile robots (AMRs) are increasingly recognised as a sustainable solution to urban challenges such as environmental pollution and labour shortages. However, dynamic obstacles in urban areas with high pedestrian density hinder their safe and efficient operation, making it increasingly important to identify favourable segments and routes for them to navigate. This study proposes the construction of a robot-specific network that incorporates link-level resistance factors, including sidewalk width and obstacle locations, as weighted attributes. Within the study area, origin–destination pairs were defined, and urban network analysis was applied to identify high-activity pedestrian segments and derive potential paths. The applicability of the proposed network was validated through field experiments using AMRs under real-world pedestrian conditions. The results confirmed that segments with frequent pedestrian robot conflicts were associated with significant travel delays, whereas paths with fewer conflicts demonstrated superior travel performance. Overall, the study provides empirical evidence supporting the necessity of robot-specific networks that explicitly incorporate pedestrian density and congestion.