Leveraging Call Detail Records to Uncover Azimuth Configuration Errors in Cellular Networks

Mobile positioning data has proven crucial in addressing many problems related to population distribution, estimating people's daily commuting, and the pandemic's spread. However, the inconsistencies and uncertainties of mobile network characteristics introduce biases and limitations in mobility pattern extraction from Call Details Records (CDR) data. Hence, understanding these factors is crucial to improving accuracy and reducing uncertainty. In this paper, we propose a new approach to detect cellplan inconsistencies, focusing on azimuth errors resulting in apparent cell-swapping. The proposed model allowed us to detect cellplan distortion by analyzing the map-matched trajectories extracted from passive mobile positioning data or CDR. The method uses only Cell Global Identity (CGI) data for positioning, without need for any additional information like Timing Advance or global navigation satellite system data. The experimental results based on actual passive mobile positioning data demonstrated the efficiency of the proposed algorithm.