Accuracy Improvement for Localization in Outdoor Area Using Artificial Intelligence

DOI: 10.14416/j.ind.tech.2021.12.005

Abstract

In this article, we propose to apply two artificial intelligence methods, Gradient descent and Adam (Adaptive movement estimation), in order to increase the accuracy in localization for an outdoor area. Both methods use an iteration to minimize error. The difference between the two methods is that Adam applies moving average to reduce error while Gradient descent only uses partial derivative. We select a True-range multilateration method to estimate a coordinate of an interesting object. Simulation results show that both Gradient descent and Adam perform approximately the same. Both methods can reduce error from the conventional method by 10.17% for a 2,500 square meters area. For complexity, we observe that the Adam method requires much less complexity than Gradient descent dose in very small area size; however, Gradient descent is slightly less complex than Adam as the area size becomes bigger.