Towards optimal positioning and energy-efficient UAV path scheduling in IoT

The Unmanned Aerial Vehicle (UAV) based communication has been emerged as a feasible solution for remote applications such as disaster management, and search and rescue due to its mobility and cost efficiency. The prior researches in this domain focused on positioning and path planning of UAVs; however, these approaches faced several limitations due to adverse weather conditions of the environment. In this paper, the impact of weather-based positioning and path planning of UAVs (IWPOP-UAV) is carried out to achieve increased QoS, reliability and energy efficiency in UAV communications. Initially, the prediction of weather conditions in the emergency situations is performed by utilizing the Cerebral Long Short-Term Memory (C-LSTM) which possesses negligible training loss and increased accuracy. The cell-based partitioning of emergency area is carried out in order to determine the target UEs. The decision upon number and position of the UAVs in each cell is provided by the A3C algorithm based on weather conditions and other significant factors thereby achieving increased coverage ratio and minimal requirement of UAV transmit power. The path planning of the UAV in order to perform effective collection of data is considered as a multi-objective optimization problem and executed by using Mayfly Optimization Algorithm (MOA). By doing so, the proposed approach is able to achieve increased QoS, reliability and energy efficiency in UAV based communication. The proposed IWPOP-UAV approach is experimented in NS 3.26 and evaluated in terms of performance metrics such as coverage ratio, cell coverage, delay, path gain, number of collected packets, UAV transmit power, and energy consumption. The results obtained are summarized and concluded to demonstrate the efficacy of the proposed approach.

Automated summary

This paper proposes an approach called IWPOP-UAV for improving QoS, reliability, and energy efficiency in UAV-based communication. It utilizes C-LSTM for weather prediction, A3C algorithm for determining the number and position of UAVs, and MOA algorithm for path planning. The approach is evaluated using performance metrics and is shown to be effective.