Anfis Based Thrust Estimation of a Small Rotary Wing Drone

Abstract

Unmanned Aerial Vehicles (UAVs) has an increasingly application for military and civilian fields. Currently, UAVs can perform manytasks such as search-rescue, surveillance by safely. UAVs are specifically designed for their using purpose. The designs of UAVs arean important and long process. It is necessary to evaluate many parameters in the thrust system design. Because of thrust system is themost important system of UAVs. Traditional thrust system design is a trial and error method that is costly and ineffective. In thisstudy, we examined that UAVs which use brushless motor in thrust system. The force generated by the thrust system has beenestimated by using the Adaptive Neuro-Fuzzy Inference System (ANFIS). In the ANFIS model, the thrust force estimation was madeby using propeller and motor information. RCbenchmark 1580 model dynamometer was used to measure the accuracy of the ANFISestimates. Mean square error (MSE) was used to compare test data and ANFIS model. Low MSE ratio shows that ANFIS model isnear to real data.