The increasing adoption of Unmanned Aerial Vehicles (UAVs) in logistics requires efficient methods for estimating their economic and environmental sustainability compared to more traditional vehicles. However, UAVs structure can vary sensitively on the market, as different models with different sizes and technical characteristics are available. This increases the difficulty of conducting data-intensive environmental analysis like life cycle assessment (LCA). This study presents a data-driven tool to predict the structure of a UAV (components and weight) based on payload specifications, addressing the challenge of UAV configuration estimation under limited technical information. A dataset of 90 UAVs was analysed, distinguishing payload-dependent components (e.g., propulsion system, frame, battery) from constant components (e.g., sensors, electronics). The predictive model was trained and validated, and its performance was assessed through error distribution analysis to detect overfitting or underfitting. Beyond UAV component estimation, this tool can be used within a decision-support tool for Life Cycle Assessment (LCA), enabling stakeholders to evaluate the environmental feasibility of UAV-based logistics before implementation.
A Data-driven model for supporting a modular bill of materials for deliveries by Unmanned Aerial Vehicles in last mile logistics
Rubrichi L.
;Gnoni M. G.;Tornese F.
2025-01-01
Abstract
The increasing adoption of Unmanned Aerial Vehicles (UAVs) in logistics requires efficient methods for estimating their economic and environmental sustainability compared to more traditional vehicles. However, UAVs structure can vary sensitively on the market, as different models with different sizes and technical characteristics are available. This increases the difficulty of conducting data-intensive environmental analysis like life cycle assessment (LCA). This study presents a data-driven tool to predict the structure of a UAV (components and weight) based on payload specifications, addressing the challenge of UAV configuration estimation under limited technical information. A dataset of 90 UAVs was analysed, distinguishing payload-dependent components (e.g., propulsion system, frame, battery) from constant components (e.g., sensors, electronics). The predictive model was trained and validated, and its performance was assessed through error distribution analysis to detect overfitting or underfitting. Beyond UAV component estimation, this tool can be used within a decision-support tool for Life Cycle Assessment (LCA), enabling stakeholders to evaluate the environmental feasibility of UAV-based logistics before implementation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
