Temporal post-processing of position estimates exploiting Markoff models

The demand for accurate positioning of mobile terminals for emergency services like E911 inspired extensive research not only on position estimation in mobile radio scenarios but also on new location based services. Unfortunately, mobile radio scenarios are characterized by shadowing and multipath propagation, which renders accurate position estimation a difficult task. This is especially true for indoor scenarios. Both position estimation based on satellite signals like GPS or GALILEO and position estimation based on the mobile radio system's signals suffer from these adverse propagation conditions. In the present paper we just accept that initial position estimates in mobile radio scenarios are rather rough. We focus on combining multiple successively obtained rough position estimates in order to obtain improved estimates of the path on which the mobile terminal moved or of the position at a certain time instant. When combining position estimates we have to consider that the mobile terminal may move during the measurement period in which the initial position estimates are obtained. We describe the mobility of the mobile terminal by a Markoff Model, whose states correspond to the positions, and apply the theory of hidden Markoff models for obtaining the improved position estimates