Research On Improved Indoor Positioning System Using UWB Signals

In recent years, high-precision positioning systems are getting more and morepeople’s attention, such as item tracking, search and rescue, battlefield soldierspositioning. GPS positioning system is the most common positioning system, butbecause of the GPS signal does not have the ability to penetrate walls, such as in acomplex indoor environments,the accuracy is limited, Ultra Wideband signal canachieve high ranging accuracy due to its ability to anti multipath and penetrate wallsThe research starts from the demand of inoor high-precision positioning.Design alow complex and low cost UWB localization system,and find a suitable solution forindoor precise positioning and verify its performance by expeimental test.The specificstudy includes the following aspects:First,The system was implemented by two way ranging protocol using TOAmeasurements.An anchor distinguish method was develop based on GML-ED(generalmaximum likelihood energy detection) algorithm, experimental data are collected toverify the the system.Second, Non line of sight (NLOS) propagation is one of the main error sources ofUWB TOA estimation. First, the direct path (DP) component may be false detected dueto the multipath dispersion in NLOS environment. Second, additional time delay ofarrival introduced by NLOS propagation will lead to a positive bias in TOA/distanceestimation. Therefore, it’s important to develop methods that can mitigate NLOS effectsduring accurate ranging and localization. Based on the features extracted from thereceived waveform in practical experiments, a machine learning method is proposed forUWB NLOS identification in this paper. Corresponding NLOS error mitigation methodis also given based on the identification results, which makes it practical in universalapplications.Finally,positioning accuracy can be further impoved through cooperativelocalizaiton instead of high-intensive anchor node configuration and high transmitpower,both of which are cost prohibitive and impractical in realistic settings.Acooperative method was developed based on iterative algorithm and compared withnon-cooperative method,simulation results show that cooperative localization can get asignificant gain in localization accuracy.