Positioning Model Training Across Uplink Power Variations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems, particularly 5G NR, face challenges in improving positioning accuracy due to the heterogeneity of uplink transmit power settings, which can degrade the learning process of positioning models.
Innovation Solution
Implementing a method where user equipment (UE) and network nodes transmit and measure positioning signals using different transmit power settings to train a positioning model, utilizing AI/ML techniques to enhance the robustness of positioning calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If positioning models are trained using positioning signals with varied transmission power settings, then the accuracy and robustness of positioning models is improved, but the device complexity and training time increase
Solution Approach 1:
The system performs preliminary actions by collecting positioning signals with varied transmission power settings in advance during normal operation. Network nodes store these signals and their associated power settings, creating a pre-prepared training dataset that can be used later for model training without requiring real-time complexity
Solution Approach 2:
The training process is segmented into distinct phases: signal collection phase where signals are gathered with different power settings, data processing phase where signals are organized and labeled, and model training phase where the actual positioning model is trained. This segmentation allows each phase to be optimized independently and reduces overall system complexity
2Reliability
If positioning signals are transmitted with multiple transmission power settings, then the robustness of location determination is improved, but the loss of time for signal transmission and processing increases
Solution Approach 1:
The system maintains continuity of useful action by collecting positioning signals with varied power settings during normal continuous operation rather than requiring dedicated training periods. Network nodes continuously receive and store these signals as they occur, ensuring that data collection does not interrupt normal positioning services or require additional time
Solution Approach 2:
Training data is collected in advance during regular system operation, so that when model training is needed, the data is already prepared. This preliminary collection eliminates the need for time-consuming real-time data gathering during actual positioning operations
Data Source
AI summary
A user equipment (UE) may receive a configuration message including a configuration for a transmission of a set of positioning signals associated with a plurality of transmission (Tx) power settings. The set of positioning signals may include a first subset of positioning signals and a second subset of positioning signals. The UE may transmit the first subset of positioning signals using a first Tx power setting of the plurality of Tx power settings. The UE may transmit the second subset of positioning signals using a second Tx power setting of the plurality of Tx power settings.