Localization of a mobile robot using indoor GPS system

In this study, an efficient ultra-wideband technology based indoor positioning system is developed to improve the measurement accuracy of a mecanum wheeled mobile robot. Data processing structure of the system is created using Monte Carlo - Latin Hypercube Sampling based machine learning algorithm. Monte Carlo - Latin Hypercube Sampling method can be used to perform the position estimation since it enables the range of each variable is fully accessible. This makes the algorithm outputs more accurate and reliable and thus increases correctness in measurements. An experimental setup including hardware, software and mathematical model is developed. A number of experiments are conducted to validate the indoor positioning system proposed. Efficiency, accuracy and precision of methodology are also tested in the experimental studies. The adaptation of the algorithm in a real system and the field tests are also conducted using a four-mecanum-wheeled mobile robot. The experimental studies show that the position estimation is improved by more than 20% using the proposed algorithm compared to the methods presented in the literature.