Integrating Artificial Intelligence in Ultra-wideband (UWB) Indoor Navigation Systems to Boost Precision and Reliability

Abstract

Ultra-wideband (UWB) technology can help with indoor guidance and finding your way around with great accuracy. This science is very important. This might be because it has a high time precision and is not affected by multipath interference. There are several things that could make it less effective than other methods, such as signal confusion, situations where there is no line of sight (NLOS), and changes in the surroundings. This is true even though it has many benefits. The latest improvements in artificial intelligence (AI) programs have made UWB-based systems much better at handling their surroundings. They are also more reliable and adaptable. Because of these improvements, UWB-based products can now adapt better to the place where they are. It is the goal of this project to create a new UWB positioning system that uses artificial intelligence to do a better job than current systems. The Deep Reinforcement Learning (DRL) and Graph Neural Networks (GNNs) will be used to make this happen. During the tests, it was found that placing accuracy rose by 37%, with an average mistake of 7.5 centimetres, even though the conditions in the setting were changing all the time. The method described works even when there is no line of sight (NLOS). This makes it a great choice for uses like self-driving robots, smart stores, and healthcare guidance.