New methods for processing GSM radio measurement data: applications for locating, handover, and network management

The GSM system provides measurement data about the radio propagation situation that is used for power control and handover decisions. Currently implemented algorithms use averages of the measurement data and comparison with thresholds for these decisions. As the variations of the measured receive levels are characteristic for the route the mobile station is travelling on, pattern recognition is proposed to determine the position of the mobile station and for handover control. Two approaches are presented-one based on neuro-fuzzy methods and the other based on hidden Markov models-and it is shown that patterns are sufficiently characteristic for the proposed application