Indoor Localization System Based on RSSI-APIT Algorithm

Area of Science:

• Computer Science
• Electrical Engineering
• Geomatics Engineering

Background:

• Indoor localization systems face challenges with accuracy due to signal fluctuations and multipath effects.
• Conventional methods like APIT struggle with precision in wide-area scenarios.
• Machine learning integration is key to overcoming limitations in current indoor positioning technologies.

Purpose of the Study:

• To develop an enhanced indoor localization system by improving data preprocessing and localization algorithms.
• To mitigate RSSI fluctuations and multipath effects using Gaussian filtering and artificial neural networks (ANN).
• To increase the accuracy and stability of indoor localization through a novel RSSI-APIT algorithm.

Main Methods:

• Fusion of Received Signal Strength Indication (RSSI) and Approximate Perfect Point-In-Triangulation test (APIT) localization methods.
• Integration of Gaussian filtering and Artificial Neural Network (ANN) for RSSI data preprocessing.
• Incorporation of RSSI ranging function into APIT for improved wide-area localization accuracy.

Main Results:

• The RSSI-APIT algorithm successfully reduced localization errors by approximately 2.9 m compared to trilateral localization and 1.8 m compared to traditional APIT.
• Localization error was consistently controlled within 1.55 m in a 100 m² complex environment.
• The system demonstrated reduced anchor call frequency, leading to lower operating costs and enhanced localization accuracy and stability.

Conclusions:

• The RSSI-APIT algorithm offers a significant improvement in indoor localization accuracy and stability.
• The integration of advanced signal processing and machine learning effectively addresses multipath effects and enhances positioning precision.
• This enhanced system provides a cost-effective and reliable solution for precise indoor positioning across various environments.