From 1G to 4G, different advances on network-based localization have been developed and included. The 3rd-Generation Partnership Project (3GPP) has being working on these standards defining localization features, such as the Positioning Reference Signals (PRS) and the Long-Term Evolution (LTE) Positioning Protocol (LPP). However, network-based localization has been always considered an optional feature for cellular networks due to its low accuracy, and its methods have been focused mainly on assistance data for GNSS and cell ID enhancement. Now, a new perspective came up in the latest releases of 4G LTE and 5G due to the introduction of high-accuracy positioning services. 3GPP is moving towards including localization for a new range of markets, which has been translated in specific 3GPP activities, aiming at providing high accuracy GNSS for LTE and 5G technologies and designing Radio Access Technology (RAT)- dependent technologies to meet more stringent targets than in previous generations. For high-accuracy positioning, for instance to support autonomous driving or industrial automation, the integration of GNSS (augmented with precise or differential corrections), terrestrial (RAT-dependent) technologies and complementary sensors is expected to play a key role on 5G localization. The goal of GINTO5G project is to support the design of PNT solutions in the context of 5G applications by carrying out extensive experiments and simulation campaigns, as well as theoretical assessment of possible disruptive techniques. For downlink TDoA using 5G SRS signal, the field trials of one campaign shows that sub-meter accuracy can be achieved with 100 MHz bandwidth in the 3.7 GHz band. At the same time the evaluation shows a significant discrepancy between achieved TOA accuracy, and the overall positioning performance, especially for the outdoor tests. Based on CEP95 and SEP95 values, it can be stated that a 2D accuracy of sub 3 meter can be achieved an outdoor area where transmitting points have been deployed and optimized for positioning purpose. Similar performance could be seen in the results of the tests carried out in indoor spaces; what is more, half of all measured indoor positions even show a significantly lower error (sub 1 meter for 2D, and sub 3 meters for 3D). Another set of outdoor trials, conducted this time on a set of transmitting points deployed more randomly, revealed a mean 2D positioning error ranging from sub metre to several hundreds of metres.