Research Information System
Neural Computing and Applications, vol.37, no.15, pp.8945-8961, 2025 (SCI-Expanded, Scopus)
In the domain of unmanned aerial vehicles (UAVs), evaluating electric propulsion systems is pivotal for enhancing performance and efficiency. This study employs a scaled conjugate gradient (SCG) algorithm to train an artificial neural network (ANN) for the propulsion system evaluation, offering a cutting-edge alternative to traditional experimental methods. The ANN architecture consists of an input layer, a single hidden layer, and an output layer. By varying the number of neurons in the hidden layer from 1 to 100, the optimal configuration with 2 neurons was identified, achieving high predictive accuracy. The model was trained using experimental datasets, predicting thrust force with an overall R2 value exceeding 0.99 across training, validation, and testing phases, and a low overall prediction error of 1.27%. These results demonstrate the ANN’s capability to generalize from training data, making it a valuable tool for UAV designers. Integrating ANN-based evaluations accelerates decision-making processes and optimizes UAV performance, marking a significant advancement in UAV technology.