Forecasting Delivery Pattern through Floating Car Data: Empirical Evidence

Comi, Antonio; Polimeni, Antonio · 2021 · Crossref

DOI: 10.3390/futuretransp1030038

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the challenge of forecasting freight delivery patterns and estimating origin-destination (O-D) vehicle flows using Floating Car Data (FCD). Traditional methods for modeling city logistics rely on resource-intensive ad hoc surveys, which often suffer from limited spatial and temporal coverage. The authors propose a discrete trip-chain order model based on random utility theory to characterize delivery tours for light goods vehicles (laden weight < 3.5 tons). The primary objective is to calibrate this model to estimate the number of stops or deliveries per tour, thereby enabling the preliminary assessment of parking demand and traffic impacts without relying on extensive manual surveys. The research utilizes a large dataset of FCD collected from 1,592 light goods vehicles operating in the Veneto region of Northern Italy over 60 working days between January and June 2018. The database comprises more than 35,000 tours, with over 8,500 specifically involving the province of Padua. The study area was zoned into 104 traffic zones for Padua and 36 zones for the surrounding provinces. The authors processed the GPS data to identify vehicle stops and classify tour patterns into four categories: single direct, multiple direct, single peddling (loading/unloading), and multiple peddling. The model formulation estimates the probability of tour characteristics, including start time, number of trips, and vehicle type, by incorporating land-use factors such as the number of wholesalers, inhabitants, and distance to service zones. Data analysis revealed that single peddling tours constituted the majority of activity, accounting for approximately 76% of all tours, followed by single direct tours at 19%. Multiple peddling and multiple direct tours represented 4.6% and 0.4% of the dataset, respectively. Temporal analysis indicated that 75% of tours began between 05:00 and 10:00, with a peak occurrence between 07:00 and 08:00. Regarding tour complexity, 24% of tours involved two trips, 20% involved one trip, and 16% involved three trips, with the average number of trips per tour calculated at 3.3. The calibrated discrete choice model demonstrated satisfactory results in estimating tour characteristics and showed effective transferability. The significance of this work lies in its demonstration that automated FCD can effectively replace or supplement traditional surveys for calibrating freight tour models. By successfully modeling the number of stops per tour, the study provides a tool for urban planners to simulate delivery impacts, estimate O-D matrices, and assess parking demand. This approach offers a more efficient and scalable method for understanding freight vehicle behavior and its interaction with urban traffic, particularly in dense urban areas like Padua.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• exposure measurement