Research And Simulation On Semi-distributed TOA-DOA Multi-node Cooperative Localization Algorithm

With the popularization and developm ent of wireless technology, wireless local network positioning service has caused much attention increasingly. The great vitality in the area of vehicle na vigation and collis ion avoidance, firefighting, ocean resc ue, unmanned aerial vehicles(UA Vs) navigation and comm unication, etc., exploited and promoted the development and application of wireless local positioning technology. Many states and resea rch institutes have attac hed importance to th e wireless local positioning technology. Among kinds of local networks, mobile Ad hoc network(MANET) has become a research hotspot. Most of the published lite ratures about the MANET were based o n the GPS to attain th e nodes’ position. However, in m any scenarios GPS service or any other exterior position services are not availab le. A self-positioning method is urgent necessary in MANET.This thesis investigates a high perfor mance cooperative localization m ethod in MANET without GPS and based on multi-node time-of-arrival(TOA) and direction-ofarrival(DOA) estimation in line-of-sight(LOS) and Non-LOS(NLOS) scenarios. Here, two categories of nodes are assumed in the MANET: base nodes(BNs) and target nodes(TNs).BNs are equipped with antenna arrays and capable of estimation TOA(range) and DOA(angle). TNs are equipped with Om ni-directional antennas and communicate with BNs to allow BNs to localize TNs. The research contents of this thesis just as follow.Firstly, a LOS localization m ethod is introduced, which is based on sem idistributed multi-node TOA-DOA fusion. We assume LOS is availab le between BNs and TNs. One BN is selected as the reference BN, and ot her nodes are localized in the coordinates of the reference BN. Each BN can localize TNs located in its coverage area independently. In addition, a TN m ight be localized by multiple BNs. Multi-n odes TOA-DOA fusion m ethod is applied to a ttain high accuracy and compared with TOA-only localization m ethod. The reference BN selecti on method and coordinate systems direction m ethod will be talked about in detail. In scenario where two BNs cannot communicate with each, localization via intermedia TN is in troduced. Federal Kalman Filter, traditional Kalman Filter and Extended Kalm an Filter are used to optimize the multi-node TOA-DOA fusion results.Secondly, in scenarios where LOS channels are not available, the precision of the proposed TOA-DOA method will not be reliable. In order to discriminate or cancel the NLOS error, a RSS data f usion model will be designed f or an error-discriminating algorithm. In this algorithm, the NL OS distance m easurements will be LOS-reconstructed and the RSS m easurements will be added to the T OA-DOA localization method.Multi-node TOA-DOA fusion m ethod and Fe deral Kalman Filter plu s fusion method are sim ulated in LOS scena rio on Matlab platform. By com paring with the traditional TOA localization m ethod and trad itional Kalman Filter, the com parison results confirm that the proposed TOA-DOA method is suitable for moderate and small scale MANET, while TOA method is suitable for large scale MANET. And, the Federal Kalman Filter plus f usion can perform more stable than traditional Kalman Filter. The NLOS fusion method can discriminate the NLOS error. The research of this thesis can provide theoretical guidance for the MANET.