نهي الناصري

Recently, the indoor location detection that utilizing Global Positioning Systems (GPS) approximation an‎d position estimation have many significant challenging issues. Due to the easiest an‎d availably of the WiFi devices an‎d mobile phones, employing WiFi signals to provide precise location detection inside a building is the subject of this work. It can be used in many applications, including military, civil, industrial, etc. The WiFi enabled device including laptops an‎d smartphones, can use the system, which uses five access points (APs) with Received Signal Strength Indicator (RSSI) measurements to pinpoint a userʹs location with fingerprinting localization system. The system is compact, energy-efficient, cost-effective, an‎d responsive. RSSI measurements were methodically recorded at predetermined reference sites throughout the floor in a real-world indoor setting in order to create a reliable fingerprinting database. The suggested IPS divides the userʹs location into ten different zones that are spread unevenly throughout a single floor. The work assesses the performance of several models for both classification an‎d regression tasks, using machine learning (ML) approaches to obtain robust indoor positioning. Decision Trees (DT), Linear Discriminant Analysis (LDA), Naïve Bayes (NB), Support Vector Machines (SVM), K-Nearest Neighbors (KNN), an‎d Logistic Regression (LR) are the six machine learning techniques used for classification. To find the best method for zone-based localization, these models are trained (offline) an‎d eva‎luated (online) using a compiled dataset of WiFi signal strengths. Three regression-based techniques are also developed to predict 2D coordinates in an indoor environment for fine-grained position estimation: Support Vector Regression (SVR), K-Nearest Neighbors (KNN), an‎d Weighted K-Nearest Neighbors (WKNN). Matlab is a system programming language that is useful for supporting ML libraries. The classification accuracy an‎d speed using several ML algorithms demonstrate good accuracy with suitable execution time, reaching up to 90% accuracy with a precision to less than 2 meters within a matter of second.

شاخص قدرت سيگنال دريافتي (RSS) , سيستم مكان‌يابي اثر انگشت , يادگيري ماشيني , نرم افزار متلب

Received Signal Strength Indicator (RSSI) , fingerprinting localization system , machine learning , Matlab program

Alnaseri

Dr. azhari