A Hybrid Learning Framework for Energy-Aware and Safe UAV Package Delivery

In the area of transportation and logistics, Unmanned Aerial Vehicle (UAV) offers certain unique advantage towards package delivery services in terms of environmental impact, operational cost, and delivery time. Review of literatures shows active participation of Artificial Intelligence (AI) is noted to be predominantly assumed as effective solution. Therefore, the proposed study introduces a hybrid approach for safer, autonomous, and cost-effective dispatching of the package from one to another location. The hybrid model uses reinforcement learning for path planning in adaptive manner while supervised learning is used for optimizing energy with payload awareness. Implemented in python, the proposed hybrid model is tested on standard dataset to find it offers 98% of mission success rate with consumption of only 5.9 minutes with minimal energy consumption in contrast to baseline methods.