A Method and Effect Exploration of Indoor Localization Technology Based on CNN and Adaptive Volume Kalman Filtering

Ultra-wideband (UWB) technology has made significant progress in indoor positioning and is widely used in various scenarios. However, due to the complexity of real-world environments, ultra wideband positioning is often affected by non line of sight (NLOS) interference from obstacles, resulting in inaccurate positioning. In order to improve positioning accuracy, this paper proposes a method that combines convolutional neural networks (CNN) and normalization operations to process raw UWB positioning data. Specifically, this article uses CNN and normalization operations to correct time of arrival (TOA) ranging data, and employs attention mechanisms to focus on key data, in order to obtain corrected ranging values. Subsequently, we used an adaptive volume Kalman filter (ACKF) to filter and track the corrected data. The experimental results showed that this method has excellent positioning accuracy, and the positioning loss converges to a small and consistent value.