User Equipment Neural Network Positioning for Indoor Accuracy

The patent applies universality by making the user equipment capable of performing multiple positioning functions - both outdoor satellite-based positioning and indoor neural network-based positioning. The device selectively activates the appropriate positioning method based on the environment, achieving both outdoor accuracy and indoor capability through a single multi-functional system.

The patent introduces wireless network signals as an intermediary for indoor positioning when satellite signals are unavailable. The neural network processes these intermediate wireless signals to derive position information, creating a mediator system that bridges the gap between satellite-based outdoor positioning and indoor positioning requirements.

The patent replaces traditional signal-based positioning mechanics with a neural network-based system. Instead of relying on complex signal propagation models and time difference calculations that limit accuracy, the neural network learns complex patterns from wireless signals and directly predicts position, achieving higher accuracy through intelligent processing rather than mechanical measurement methods.

The patent changes the processing parameters by using a neural network to analyze multiple wireless signal parameters simultaneously. The neural network processes signal strength, signal quality, and other wireless signal characteristics as input parameters to determine position, transforming the approach from single-parameter measurement to multi-parameter intelligent analysis.

The patent applies preliminary action by pre-training the neural network model with positioning data before deployment. The neural network is trained offline with labeled position and wireless signal data, so that during actual operation, it only needs to perform inference rather than learning from scratch. This preliminary training phase separates the complex learning process from the operational phase, reducing real-time processing complexity.

The patent implements self-service by enabling the user equipment to autonomously perform positioning using its own neural network model and locally processed wireless signals. The device independently determines its position without requiring complex centralized processing or extensive external infrastructure, making the system self-sufficient and reducing overall system complexity.