Improvement of fingerprinting technique for UWB indoor localization

The indoor localization is highly desired for elevation in the industrial, public safety and medical technology. Its significant requirement is a high accuracy in dense multipath fading environments. This paper improves the main problem of indoor localization by using fingerprinting technique based on ultra wideband (UWB) channel measurement. In addition, the neural network algorithm is used to find the location that has 3 dimensions, x plane, y plane and z plane. The first path loss, the delay times of the first part, the average path loss and the average of delay time were applied to create the ultra wideband radio propagation parameters for fingerprinting technique. the results are shown in terms of histogram of distance errors between update and non-update the databases and countour graph of each of distance errors on the coordinates. According to the results, the fingerprinting technique with update database can identify the location more correctly than the fingerprinting technique with non-update database. It had accuracy at 1 meter as 88.57%. while the fingerprinting technique with non-update database was affected from the environment and the accuracy at 1 meter was lower than the fingerprinting technique with update database all the times of experiment.