Enabling Green Crowdsourced Social Delivery Networks in Urban Communities

Choi, Kevin; Bedogni, Luca; Levorato, Marco · 2022 · Crossref

DOI: 10.3390/s22041541

archive: archived pipeline: cataloged verified

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

Summary

This paper addresses the inefficiencies and environmental costs associated with the "last mile" of urban delivery, which accounts for approximately 28% of total delivery costs and contributes significantly to traffic and pollution. The authors propose a novel model for "Green Crowdsourced Social Delivery Networks," which leverages existing transportation flows from pedestrians, cyclists, and drivers to deliver packages without introducing additional vehicles or routes. By utilizing data from wearable devices and GPS traces, the study aims to create ad hoc delivery networks that are zero-emission, cost-effective, and capable of accessing areas difficult for traditional logistics, such as car-free zones or indoor environments. To evaluate the feasibility of this model, the researchers developed a custom simulator that matches delivery requests with available carriers based on minimizing both delivery time and the carrier’s deviation from their original route. The study utilizes two distinct datasets: a large-scale dataset of New York City taxi traces (representing a dense, motorized scenario) and a smaller dataset of pedestrian and bicyclist routes from the University of California, Irvine (UCI) campus (representing a constrained, non-motorized scenario). The NYC data, comprising over 12 million taxi journeys, was subsampled to match the population density of the UCI dataset to allow for comparable analysis. The simulator employs metrics such as Capacity Ratio, Success Ratio, and Delivery System Efficiency Rating (DSER) to assess performance, accounting for user willingness to wait and carrier willingness to divert. The results demonstrate that crowdsourced delivery networks are viable and can significantly reduce delivery times and emissions. In the New York City simulation, the system maintained a 100% success rate even when handling 10,000 requests, saving users an average of 8.96 minutes and reducing total driving time by 10.78 minutes per delivery. The study found that efficiency metrics degrade as the number of requests increases, but the system remains effective within reasonable capacity limits. Furthermore, the authors extended their work to predict future carrier routes based on historical trace data, showing that route predictability is feasible. The findings were consistent across different city subsets and carrier types, indicating that the model can be generalized to various urban environments. The significance of this work lies in its demonstration that leveraging existing human mobility patterns can create sustainable, efficient last-mile delivery solutions. By avoiding the need for dedicated delivery infrastructure like lockers or additional vehicles, this approach offers a scalable solution for dense urban areas and campuses. The paper concludes by laying the groundwork for future real-world deployments, suggesting that such networks can effectively reduce traffic congestion and carbon footprints while providing faster service to end-users.

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.

• last mile delivery