| The rapid development of the internet of things(Io T)and the emergence of a large number of smart wireless terminal devices have promoted the rapid development of the new ecosystem of smart Io T,and expanded new landing sites and application scenarios for short-range wireless communication technologies such as wireless sensor networks(WSN)and wireless local area network(WLAN).WSN has the characteristics of wide distribution,large number of nodes,low cost of a single node,strong network dynamics,easy expansion and fast networking.It has been widely used in military,environmental testing,medical care,smart home,smart agriculture and other fields to achieve monitoring and management.WLAN has the advantages of low hardware cost,convenient installation and deployment,easy expansion,and fast networking.It has been widely used in airports,shopping malls,schools,office buildings and other places to realize wireless access to the Internet.In the Io T and industrial internet systems,there are more and more location-based services,which further promotes the development of positioning technology.As an important feature of the physical world,location information is a basic element for providing location-based services in the Internet of Things system.Positioning technology can provide important support and guarantee for communications and services that relying on the location information of objects and users in the Internet of Things.In the open outdoor environment,global positioning system(GPS)can provide accurate positioning services.At present,positioning systems based on WSN and WLAN have been widely used in the Internet of Things and Industrial Internet to achieve positioning services in indoor or complex outdoor environments where GPS cannot achieve good positioning results.At present,positioning systems based on WSN and WLAN have been widely used in the Internet of Things and Industrial Internet to achieve positioning services in indoor or complex outdoor environments where GPS cannot achieve good positioning results.The positioning problem in the scenario of Internet of Things and industrial Internet is the positioning problem in the scenario of multi-layer network deployment.Aiming at the positioning problem of the multi-layer WSN structure and the multi-layer WLAN structure, this thesis divides it into two problems: the acquisition of plane coordinates and the acquisition of effective position information in the profile.At present,positioning technologies can be divided into range-based positioning technologies and range-free positioning technologies.The range-based positioning algorithms using time of arrival(TOA),time difference of arrival(TDOA)and angle of arrival(AOA)have the advantages of high positioning accuracy,but them have high requirements on hardware and high energy consumption.The range-based positioning algorithms have low positioning accuracy,but them have low requirements for node hardware,low positioning cost,and better robustness to environmental interference.For location-based service applications that do not require high accuracy,range-free wireless network positioning technology is more suitable.For the WSN and WLAN range-free positioning problems in a single-layer plane,researchers have solved the key technical problems in the positioning process from different aspects,but there is still research points for further improvement.In the range-free positioning algorithm of WSN,there is a problem that the positioning accuracy is greatly reduced by the hop distance instead of the straight line distance.In the single-layer WLAN indoor positioning problem based on received signal strength(RSS)fingerprints,there are problems of reduced positioning efficiency and positioning accuracy caused by huge fingerprint data and redundant fingerprint information.In the RSS fingerprint positioning problem of the multilayer WLAN structure,there is a problem that the effective longitudinal position information in the profile is inaccurate.This dissertation studies the RSS range-free positioning problem of multi-layer WSN deployment and multi-layer WLAN deployment,and proposes several innovative methods.The main innovations of this article are as follows:(1)The 3DHDCDV-Hop positioning algorithm based on hop distance correction and improved differential evolution algorithm is proposed,which solves the problem of the cumulative error introduced by hop distance instead of the distance between nodes in the RSS range-free positioning technology based on hop distance estimation in WSN.Through the analysis of radio irregularity model(RIM),the radiation range of the node is divided into two parts,the fractional hop count is introduced.The anchor node is divided into the nearest anchor node of the unknown node and other anchor nodes of the unknown node,and the average hop distance of the unknown node relative to different anchor nodes is obtained.The error introduced by solving the distance between the unknown node and the anchor node by using the uniform average distance per hop is corrected.In the unknown node position solution stage,the differential evolution(DE)algorithm based on the optimal solution mutation strategy of the population is introduced.At this stage,the evolutionary trend of the algorithm is controlled by the hop distance,and the effective longitudinal position information in the multi-layer structure profile is used to further limit the evolutionary trend,so as to alleviate the influence of distance accumulated error on the positioning result.The simulation results show that the proposed algorithm can improve the positioning performance of the original positioning algorithm based on hop distance estimation,and by setting a reasonable communication radius and anchor node ratio,the node positioning performance and network cost can be balanced.(2)A fingerprint positioning algorithm based on database partition processing is proposed,which solves the problem that the single-layer WLAN structure reduces the location performance due to the large amount of fingerprint data and the redundancy of fingerprint data.A three-level fingerprint database processing method is proposed.The proposed density cluster clustering method based on adaptive centroid acquisition is used for the first level partition processing of the fingerprint database.The principal component analysis(PCA)method is used for the second level dimension reduction processing of the obtained sub fingerprint database.The k-d tree algorithm is used for the three-level hierarchical storage of the reduced sub fingerprint database.In the stage of online location determination,the weighted similarity measure is used instead of the fingerprint distance measure to determine the subregion where the online fingerprint is located,which avoids the one-sidedness of using the central fingerprint data in the subfingerprint database to determine the subregion where the online fingerprint is located.Simulation results show that the proposed algorithm can improve positioning performance.(3)A method for identifying effective longitudinal position information based on access point(AP)selection and multi-class linear discriminant analysis(LDA)is proposed,which solves the adverse effects of the lack of effective longitudinal position information in the multi-layer WLAN structure on positioning.In the offline stage,the offline AP selection factor is used to obtain the offline effective AP set,and the multi-class LDA is used to classify and train the fingerprint information of the effective AP set on different floors to obtain the floor discrimination model.The maximum likelihood estimation method is used to obtain the Gaussian density function of the fingerprint information of the effective AP set on different floors in the projection space.In the online stage,the online AP selection factor is used to select the online effective AP set,and the same set as the offline effective AP set is extracted for floor discrimination.The simulation results show that in the case of a multi-layer WLAN structure,the probability of the proposed algorithm to obtain accurate longitudinal position information in the profile reaches 98%.In this dissertation,the problem of range-free positioning in multi-layer network scenarios is transformed into two problems: solving plane coordinates and determining effective longitudinal position information in multi-layer structure profile,and the corresponding solutions are proposed,which can provide theoretical reference for the application of wireless network positioning technology in the Internet of Things and industrial Internet. |
PDF Full Text Request