Abstract—Numerous applications within the realm of Internet of Things (IoT) involve with tracking personnel and merchandise whose quality is influenced by indoor location precision. Pattern matching of the signal approach, also referred to as the feature fingerprinting method, is one of numerous indoor positioning methods. Achieving precision in positioning is easily disrupted due to the presence of noisy ambient circumstances. Efficient stable techniques are needed to mitigate these negative effects on the localization quality. This study introduces several novel machine learning methods and indexing approaches aimed at enhancing the accuracy of indoor localization applications. Genetic algorithms and partial least square theories are proposed to work together for this purpose. Traditional fingerprint localization methods such as Particle Swarm Optimization (PSO), Gaussian models are also tested for verification purpose. This method fine tunes the major frequencies/amplitudes of the Gaussian models via the PSO algorithm trying to approximate the noisy spectrum of the Received Signal Strength Indicator (RSSI) signals. The Genetic Algorithm (GA)/Partial Least Squares (PLS)/K-Nearest Neighbors (KNN) method, when compared to the PSO/Gaussian model fingerprint methods, can achieve a precision of 92% indoor localization precision, while requiring minimal development time. In a complex laboratory and hallway setting, the total accuracy rate can reach 95% with a resolution of 16 cm when a weighted KNN algorithm is included to the target location verification procedure. Overall, our suggested GA/PLS/KNN method outperforms traditional methods and current static localization methods based on many wireless techniques such as Wifi, 4G/5G, Bluetooth Low Energy (BLE), etc.
 
Keywords—Internet of Things (IoT) localization, Particle Swarm Optimization (PSO) algorithm, Partial Least Squares (PLS) algorithm, Genetic Algorithm (GA), intelligent positioning


Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).