Since 2016, when the Google announced the availability of Raw Global Navigation Satellite System (GNSS) Measurements from the Smartphone, more than 1000 research papers have been published with the performance analysis and possible future advancement in GNSS prospective . Big industrial giants like Broadcom also came up with dual frequency chip to keep the competitors on their toes. But, to cope up with the technical advancement like autonomous driving, decimeter level of accuracy is not enough. In order to have full autonomy, high level of integrity and continuity is expected from a navigation system. A crucial function for automated vehicle technologies is accurate localization. Lane-level accuracy is not readily available from lowcost GNSS receivers because of factors such as multipath error and atmospheric bias. In recent years there have been an extended demand of the precise system for vehicular navigation. According to a recent survey by used car retailer Carzoos, 60 % of respondents would rather use a mobile device as their navigation system, compared to the 22 % who would prefer a built in GPS. With the series of experiments, it is evident that the modern smartphones with dual-frequency GNSS receivers under very low multipath environment, is capable of providing cm level accuracy   . Experiments like zero-baseline retransmission and hardware level of multipath mitigation using choke ring platform, multipath error can significantly suppressed. However, these experiments were carried out under optimal open sky conditions and with good satellite constellations. The goal of our experiment is to investigate, how the absolute position error behaves under suboptimal measurement conditions (for example within vehicles with large shadowing and less good satellite constellations). In this paper the point errors are examined in various measurement scenarios. Various options for multipath suppression and smartphone alignment and position in the vehicle are tested. Stochastic models can then be derived from the distribution of the point errors in the measurement series. For most applications, the 2D point error is assumed to be normally distributed, but this is not (always) the case. Generalized Pareto distributions often model the point error significantly better, especially for the peripheral areas. Composited distributions from several basic stochastic distributions are also conceivable. With the help of so-called over-bounding, conservative probability distributions can be generated, which make a more pessimistic statement about the point errors. The conservative stochastic models can then be used to estimate probabilities for critical point errors, which are particularly important for safety-relevant systems. A possible application here is the lane-precise localization of vehicles and the resulting error rate. The result of this paper is supposed to be a conservative stochastic model for the 2D point error for SPP. RTK-Float and RTK-Fix positioning results with modern smartphone. Finally, the model can be used to determine the integrity of location-based services.