مريم عليرضايي بافقي

In recent years, indoor positioning systems have gained significant attention due to the increasing demand for location-based services in indoor environments. Among various methods, Bluetooth Low Energy (BLE) has become one of the most popular technologies in this field due to its low power consumption, cost-effectiveness, and compatibility with smart devices. However, several challenges such as low accuracy in certain environments, negative impacts of environmental conditions, and signal interference affect the performance of these systems. This thesis investigates the use of machine learning techniques to improve the accuracy and efficiency of BLE-based indoor positioning systems and proposes new models to address the existing challenges. In this research, a review of indoor positioning technologies and common algorithms is conducted, and the main challenges in the application of BLE are examined. Subsequently, models are developed using various machine learning techniques, including Support Vector Machines (SVM), Decision Trees (DT), K-Nearest Neighbors (K-NN), and Artificial Neural Networks (ANN), to enhance the accuracy and efficiency of indoor positioning systems. These models are trained using data collected in various environments, and their results are eva‎luated. The experimental results demonstrate that the use of machine learning techniques, especially the proposed hybrid method combining the Random Forest and K-Nearest Neighbors algorithms, can significantly improve the accuracy of BLE-based positioning systems. The proposed model has shown better management of signal interference and reduced positioning errors, providing higher accuracy compared to conventional machine learning methods. Finally, this thesis offers recommendations for further improvement of these systems in complex environments and suggests new research directions for future studies.

موقعيت‌يابي داخلي , بلوتوث كم‌انرژي , يادگيري ماشين , دقت موقعيت‌يابي , الگوريتم K نزديك‌ترين همسايه , تداخلات سيگنالي

Indoor Positioning , Bluetooth Low Energy (BLE) , Machine Learning , Positioning Accuracy , K-Nearest Neighbors (K-NN) , Signal Interference

maryam alirezaei bafghi

Dr. Saeed Ebadollahi