Research And Application Of Location-awareness Algorithms In Internet Of Things Environment

With the development of science and technology,the demand for applications is overgrowing.Among these technologies,wireless location-aware technology has shown high activity in military and civilian applications.Meanwhile,wireless location-aware technology and wireless location-aware services play an increasingly important role in people's lives.Among the outdoor location-aware technologies,the satellite-based location-aware system,such as the global positioning system,is a representative location-aware technology with low cost and high stability and is widely used in military and civilian applications.With the development of Internet of Things(Io T)technology,indoor location awareness technology has attracted more and more attention.However,due to the influences,such as heterogeneous sensing device,environmental stability,obstacle obstruction,sensor deployment cost,and sensor stability,there are still some problems to be solved in practical applications.In this thesis,we study the problem of location awareness in Io T environment and focus on solving three key issues: "high-cost deployment with labeled tags,poor environmental adaptability" and "poor adaptability of sensor fault" in two-dimensional Io T location-aware algorithms,and "great deployment difficulty,low accuracy" in three-dimensional locationaware algorithms,aiming to improve the usability of the Io T location-aware algorithms in 2D and 3D environments.Meanwhile,we expand the existing reconfigurable Io T middleware of the research group and combine with the proposed location-aware algorithms to form a lightweight middleware.It can meet the needs of location awareness in the Io T environment,which provides supports for the collection of location information for the intelligent Internet of Things.The main research contents and innovations of this thesis are summarized as follows:(1)In this thesis,we fuse the glowworm swarm optimization algorithm with the semisupervised online sequential extreme learning machine to propose an environmental adaptive positioning algorithm based on radio frequency identification.The algorithm automatically adjusts the regularization coefficients of the semi-supervised online sequential extreme learning machine through the glowworm swarm optimization algorithm.Then it can obtain the optimal regularization coefficients under different initial conditions.Simultaneously,the semisupervised characteristic of the algorithm can reduce the number of labeled reference tags and the cost of positioning systems.Besides,the online learning phase of the algorithm can continuously update the system to perceive changes in the environment and reduce the interference of environmental changes on the algorithm.The experimental results show that compared with other algorithms,the proposed algorithm can achieve more accurate positioning results and continuously learn about the environmental changes,for example,in realistic experiments,the average positioning accuracy improvement rate under changing environments is 22.82%.(2)We propose a radio frequency identification reader-fault-adaptive localization algorithm based on the online sequential fuzzy broad learning system.The algorithm improves the fuzzy broad learning system with the ability of online sequential learning to process data streams that continue to arrive in the environment.Meanwhile,the algorithm proposes a readerfault-adaptive strategy,which can process subsequent data streams when the reader fault occurs and reduce the impacts on the positioning accuracy of the original system.We have carried out experiments to study the influence factors and validate the performance.The simulation and realistic experiment results show that our proposed algorithm can achieve a better positioning effect and maintain relatively high accuracy in the reader-fault environment,for example,in realistic experiments,the average positioning accuracy improvement rate under the reader-fault environment is 22.74%.(3)We propose a radio frequency identification three-dimensional positioning algorithm based on adjustable signal strength.Firstly,we deploy the reference tags to a wall or ceiling,obtain a set of candidate reference tags by adjusting the transmit power of the reader and estimate the distance between the reader and the target tag by signal strength similarity.Then,we use the sphere trilateration method to estimate the target tags' location in three-dimensional space to obtain a set of candidate positions.Finally,we use the plane fitting method and the least-squares method to optimize the candidate locations to reduce errors and obtain more accurate results.The simulation and testbed experiment results have shown that the proposed method can reach a higher positioning accuracy and more robust to the environment than other existing methods,for example,the average positioning error in a liquid metal mixed sheltering environment has increased by 0.58 meters.(4)We expand the reconfigurable Io T middleware of the research group and apply it in combination with the proposed location-aware algorithms.It forms a lightweight middleware that can meet the location-aware needs in the Io T environment.Also,we give the overall architecture and implementation of the middleware.The extended location-awareness Io T middleware consists of the localization data fusion module,the location-aware algorithm module,and the WEB interaction module.It can quickly improve the existing middleware according to different location-aware algorithms.Simultaneously,the middleware integrates the specific implementation of the previous two-dimensional and three-dimensional positioning algorithms.The experimental test results show that the extended middleware can achieve the expected effect and provides an essential reference for the location-aware algorithm application in the Io T environment.