Location of mobile terminals with quantized measurements

This paper introduces a new method for locating mobile terminals for wireless networks in dense urban locations based on low-cost quantized propagation time measurements. The joint probability density functions (PDFs) of the mobile terminal location and measurements are approximated as Gaussian mixture models based on survey data collected in the network area. Optimization is performed to find a Gaussian mixture model which describes the survey set with a minimum number of bits. This reduces the effects of over-fitting of the survey data on the approximate PDF. The approximate joint PDF of terminal location and measurements is used to estimate the location of mobile terminals and the covariance of the location error. It is shown how this error covariance can be used with Kalman filtering to combine past location estimates with the current location estimate and further reduce location error