Optimizing Indoor Positioning in Large Environments: AI

Abstract

Indoor positioning system (IPS) has been a consistent challenge for many years in real-world applications with these environments often extremely large and complex such as airports, hospitals and industrial sites that required at least one-meter accuracy. Although technologies such as Wi-Fi Fine Time Measurement (FTM), Ultra-Wideband (UWB), and 5G have been developed and show significant progress, scalability and precision are frequently impaired by a dynamic environment with numerous obstacles. This study proposes an innovative AI-driven sensor fusion technique that harnesses the power of machine learning models to fuse data from potential positioning technologies, boosting the accuracy and adaptability of the system. Real-time sensors adjust inputs as the environment changes, so proper positioning can be done much quicker than traditional methods and lead to a more efficient robot. These works also provide results of a comprehensive assessment of the system performance in different close-to-reality applications, discussing their possible implementations in the IoT context.