Résumé : Fifth generation (5G) networks are expected to provide high precision positioning estimation utilizing mmWave signals in urban and downtown areas. In such areas, 5G base stations (BSs) will be densely deployed, allowing for line-of-sight (LOS) communications between the user equipment (UE) and multiple BSs at the same time. Having access to a plethora of measurement sources grants the need for optimal integration between the BSs to have an accurate and precise positioning solution. Traditionally, 5G multi-BS fusion is conducted via an extended Kalman filter (EKF), that directly utilizes range and angle measurements in a centralized integration scheme. Such measurements have a non-linear relationship with the positioning states of the filter, giving rise to linearization errors. Counter to the common belief, an unscented Kalman filter (UKF) will fail to totally eradicate such linearization errors. In this paper, we argue that a de-centralized integration between 5G BSs would fully avoid linearization errors and would enhance the positioning performance significantly. This is done by fusing position measurements as opposed to directly fusing range and angle measurements, which inherently leads to a linear measurement model by design. The proposed de-centralized KF method was evaluated in a quasi-real simulation setup provided by Siradel using a real trajectory in Downtown Toronto. The experiments compared the performance of de-centralized KF integration to that of centralized EKF and UKF integration schemes. It was shown that the proposed method was able to outperform both UKF and EKF implementations in multiple scenarios as it decreased the RMS and maximum 2D positioning errors significantly, achieving decimeter-level of accuracy for 90.3% of the time.