Research On Localization Optimization Technique With The Ubiquitous Wireless Signal

With the rapid development of wireless communication and information technology, the commercial scale of localization technology has been increasing. The GPS-based localization technology is often unavailable in the indoor, underground, dense urban areas or under the bad weather conditions, so the land-based ubiquitous wireless signal localization becomes the necessary additional technique. Although it has been commercialized, its performance is limited by the measurement error in the complex environment, which causes low localization accuracy. At the same time, considering the contradiction between the localization accuracy and the complexity, the error mitigation technology with low complexity becomes the research hotspot in the field of localization.Through the review of the related literature and research status, three major research directions are studied in this paper:the Non-Line-of-Sight error mitigation technology, the localization technology utilizing the signal reflected paths in the multipath transmission and the node localization technology in the wireless sensor network. The research and innovation points are as follows.Regarding to the Non-Line-of-Sight error mitigation technology, the wireless localization algorithms based on distance geometry are summarized and analyzed, and it is applied to the cellular network environment. The constraint conditions based on theory of distance geometry using4and3base stations to locate are deduced respectively, and the measurements are optimized with the filter algorithm based on the data search and the constraint algorithm based on nonlinear optimization respectively. According to the characteristics of the two methods, analysis of the application scenarios is carried out. Simulation results show its effectiveness.Regarding to the localization technology utilizing the signal reflected paths in the multipath transmission, the LPMD (Line of Possible Mobile Device) algorithm utilizing the single-bound reflected path is improved firstly. And then, based on the ideology of ray tracing and dead reckoning, the signal path reckoning method with the assistance of geographic information system is proposed to solve the problem with the multi-bound reflected path. With the idealization of the building models, the signal propagation path is reckoned using the measurements of emission angle and propagation distance, and then the location estimation is obtained. Simulation shows that the proposed algorithm achieved good results when the angle error is small.Regarding to the WSN (Wireless Sensor Network) localization technology, the estimation of distribution algorithm is introduced into the localization field. Firstly, sample points are taking randomly in the intersect area of the communication range of the anchor nodes. Secondly, the better samples are chosen according to the distance error calculated. Then the new samples group is generated with the help of the probability distribution model. Finally, the better location estimation is obtained through continuous screening. Simulation results show that the proposed algorithm achieves the comparable performance with other up-to-date complex localization algorithms at low noise level.