An ANFIS Approach to Modeling a Small Satellite Power Source of NASA

Before launching satellites into space, a wide variety of practical and comprehensive tests must be done on their different subsystems. Because, the cost spent for designing and manufacturing satellites is much higher than these investigations. One of the prominent sectors of these devices is their power supplies. In this paper, a neuro-fuzzy based black-box technique for modeling a li-ion battery used in a small satellite of the National Aeronautics and Space Administration (NASA) is presented. The dataset was extracted from a range of particular tests on 18650 lithium-ion cells by scientists of NASA. The proposed approach includes an Adaptive Neuro-Fuzzy Inference System (ANFIS) model with a Fuzzy Inference System (FIS) generated by a subtractive clustering algorithm to estimate and predict the capacity of the cell for next cycles. The results indicated the proposed method can be considered an efficient and reliable technique for estimating parameters of batteries.