A new method for positioning of mobile users by comparing a time series of measured reception power levels with predictions

Many modern applications for mobile phones are based on information about the exact position of the user. There are several methods, which fulfill this task. Some require extensions of the network or the terminals; others do not, but provide little accuracy. The method presented here uses exclusively standard measurements, network parameters, and data from the network planning process. The power level measurements of the mobile reports to the network for handover and other radio resource management tasks are compared with the field strength values predicted by network planning tools. Using a single report is a well-known positioning method, which is not very accurate, because it is very sensitive to fading and other random variations of the signal levels. In the new method, a series of several reports is used. This series corresponds to a path, on which the user moves. An efficient algorithm calculates the probability of the most probable path leading to each point in the search area. The estimated position is the weighted average of all these points. Since searching the whole area of a network is not feasible in practice, the calculation effort is drastically reduced by a preprocessing step, which allows the preselection of the search area based on information about the current service cell of the user. Extensive field trials in existing GSM networks have shown that the method works well and gives substantial improvement compared to other positioning methods, such as CI/TA. An average accuracy of 130m was observed in an urban environment and about 300m in rural areas. An UTRAN provides some measurements, which are very similar to those of GSM, it is possible to apply this method in UMTS networks as well.