The performance of Global Navigation Satellite System (GNSS)-based positioning techniques degrades in tunnels, urban canyons and other areas, in which GNSS coverage is poor. Recent advances indicate that radio-based positioning techniques have the potential of complementing GNSS-based positioning services in such problematic areas. In this work, we propose to combine range and angle measurements routinely exercised by fifth generation of mobile communication (5G) base stations with acceleration measurements by vehicles to generate position estimates. Specifically, we propose to utilize an extended Kalman filter to combine the range, angle and acceleration measurements. The accuracy of this positioning system is studied in millimeter-wave 5G cellular networks. Specifically, the proposed positioning system using 5G and sensor fusion is tested in a highway scenario in which 5G base stations provide cellular coverage. In this setting, we study the impact of certain parameters of the 5G network such as the available bandwidth and the propagation environment on the positioning accuracy. We find that the positioning accuracy is largely affected by the number of antennas in the base station and that the proposed scheme outperforms GNSS based schemes in the problematic areas.