Last-Mile Delivery Optimisation with Autonomous Vehicles and Drones for Trust-Aware Multimodal Medical Logistics in Urban Environments

This study addresses last-mile urban medical logistics through a multimodal fleet of autonomous vehicles and drones, including autonomous vans and trucks. We propose an environment-aware optimisation framework for last-mile delivery optimisation that integrates trust quantification, adaptive algorithm selection, and real-time decision-making. The system models road and environmental conditions using multi-source data and selects path-planning strategies, such as Dijkstra, A* or Greedy Best-First Search, according to delivery urgency and context. A trust evaluation mechanism fuses subjective perceptions with objective performance to embed trust scores into the optimisation process. Machine-learning models predict delivery delays and guide vehicle scheduling under dynamic urban logistics network conditions. In simulations based on Coventry, UK; integrating trust metrics with environment-sensitive routing reduces mean delivery time by 28%, while delay-prediction error drops to ±3.04 s; drones deliver insulin 90% faster than trucks despite slightly lower trust scores (-1.6%). These findings underscore the value of incorporating socio-technical factors, especially trust in time-sensitive, safety-critical multimodal urban logistics.