Multimodal Autonomous Last Mile Delivery System Design and Application

TL;DR

This paper proposes and evaluates three autonomous last mile delivery systems, including a hybrid robot-drone approach, demonstrating significant efficiency improvements in urban food delivery simulations.

Contribution

It introduces a hybrid delivery system combining aerial and ground autonomous vehicles with a hub-and-spoke network, evaluated through agent-based simulation.

Findings

Hybrid system outperforms individual robot and drone systems

48% reduction in delivery time with optimized fleet

Simulation results support hybrid approach effectiveness

Abstract

With the rapid increase in congestion, alternative solutions are needed to efficiently use the capacity of our existing networks. This paper focuses on exploring the emerging autonomous technologies for on-demand food delivery in congested urban cities. Three different last mile food delivery systems are proposed in this study employing aerial and ground autonomous vehicles technologies. The three proposed systems are: robot delivery system, drone delivery system and a hybrid delivery system. In the hybrid system the concept of hub-and-spoke network is explored in order to consolidate orders and reach more destinations in less time. To investigate the performance of the three proposed delivery systems, they are applied to the city of Mississauga network, in an in-house agent-based simulation in MATLAB. 18 Scenarios are tested differing in terms of demand and fleet size. The results show that the hybrid robot-drone delivery system performs the best with a fleet side of 25 robots and 15 drones and with an average preparation and delivery time less than the individual robot and drone system by 48% and 42% respectively.