An Improved Neural Network Algorithm for the Deployment of a UAV-Aided IoT Data Collection System

In recent years, adopting unmanned aerial vehicles (UAVs) to collect the data of Internet of Things (IoT) devices has been a hot topic. Considering the limited energy of IoT devices, how to effectively deploy the UAV to optimize the energy consumed by the UAV and IoT devices is a very practical research project. To address this issue, this paper designs an improved neural network algorithm (INNA) to simultaneously determine the number and locations of stop points of the UAV. The core idea of INNA includes a random transfer strategy and a random dropout technology. The two strategies not only enrich the diversity of the population but also enhance INNA's ability to escape from local optima. Experimental results show the excellent performance of INNA.