Optimization Of UWB-based TDOA Localization

Indoor localization becomes more and more important as the development of warehousing and delivery.In indoor environments,the traditional GPS localization system has poor accuracy since the signal is blocked by the wall,which makes GPS unsuitable for these scenarios.To overcome this flaw of traditional GPS-based localization,other radio localization technics such as WIFI,Bluetooth or Ultra-Wideband(UWB)systems have been developed for GPS-denied environments.Due to advantages in high localization accuracy and low power consumption,UWB localization attracts a lot of attention,especially in industrial fields.Among all UWB localization technologies,the Time Difference of Arrival(TDOA)approach is particularly suitable for massive sensor systems because TDOA-based systems can support more target nodes than other technologies.However,TDOA-based systems are sensitive to the environmental disturbances.The accuracy of TDOA-based systems will decrease if there is a blockage between the transmitter and receiver.In addition,inappropriate selection of reference nodes can also result in low accuracy.These disadvantages somehow limit the application of TDOA localization systems.To solve the problems mentioned above,this paper proposes a refined TDOA localization system,focusing on improving the performance of traditional TDOA-based localization technique for massive sensor networks in complicated indoor environments.The main contributions are as follows:1)A TDOA-based localization system is established,which can support multi target for localization.2)To overcome the deflect of accuracy decrease in complicated environments by traditional TDOA systems,we first analyze the influence on distance and angle measurement from blockages through experiments based on a Decawave PDOA-kit in Not Line of Sight environments.Based on the experiment results,an improved Extended Kalman Filter is proposed,which fuses TDOA and Angle of Arrival measurement.Compared with traditional TDOA localization systems,the proposed Extended Kalman Filter algorithm can effectively improve localization accuracy in the Not Line of Sight environment,which is verified through simulations.3)In order to improve the relative localizing accuracy in dynamic multi-sensor systems,the theorical property of reference nodes’ selection is specifically analyzed and a dynamic reference node selection algorithm is proposed.The algorithm can update the selection of reference nodes and improve the robustness of the localization system in a massive sensor network.Simulation results validate that the algorithm could improve the localization accuracy in dynamic multi-sensor systems.