Research On RSSI Indoor Environment Perception And Location Technology Based On Machine Learning

In recent years,the development of artificial intelligence and Internet of things is developing rapidly.The emergence of some new mobile devices has stimulated the surge of location aware services.Indoor wireless location technology has gradually become the focus of attention.Therefore,how to achieve indoor location technology with adaptive,universal and high accuracy in the dynamic indoor location environment is the core of current research.The existing indoor location technology based on Received Signal Strength Indicator(RSSI)has the advantages of simple equipment and convenient data acquisition.At the same time,machine learning has the ability to learn.It can train the network by collecting sample data,add predictive function to the network,and collect different training data sets in different environments,so that the network can adapt to different environments.Based on the above ideas,it is of great practical significance to study the indoor positioning technology combined with the two technologies,so that they can learn and adapt to different positioning environments,so as to achieve high accuracy positioning effect.First of all,this paper introduces the basic knowledge of indoor positioning technology and machine learning,as well as the application potential of indoor positioning technology in various fields at the present stage,and analyzes the difficult problems faced by indoor positioning technology at present,mainly aiming at the two key technologies: adaptive indoor location environment and ranging location algorithm.Secondly,the advantages and disadvantages of existing indoor location technology based on RSSI ranging and machine learning are analyzed,and the research goal of this paper is proposed.The existing indoor location technology based on RSSI ranging usually uses the logarithmic path loss model,which involves two environmental parameters.The communication between anchor nodes can achieve real-time correction of the environmental parameters in the static indoor positioning environment,and the electronic map can be generated quickly by using the anchor nodes deployed in advance.However,it is difficult to update the two environmental parameters in the scene of dynamic change of indoor location environment caused by changes in personnel status.Aiming at this problem,a clustering algorithm based indoor activity perception and detection is proposed in this paper.The algorithm uses clustering technology to cluster the RSSI values and corresponding variance values of different personnel status collected in the environment.The thresholds of the RSSI value and the variance value in each personnel state are obtained,and the threshold is used for classification in the actual positioning,realizing the detection of the personnel state in the dynamic indoor positioning environment,reducing the influence of the personnel factor,and providing a basis for selecting the ranging and location model according to the personnel status.Finally,aiming at indoor location algorithm,it points out that the existing technology has low precision and can not adjust its own model with personnel status.An indoor location algorithm based on joint adaptive enhancement strategy is proposed.The algorithm collects data from two states of static and dynamic of the personnel to train the network firstly.After several iterations,the ranging positioning model under different personnel states is obtained,the personnel state perception and detection algorithm based on clustering is used to detect the personnel status,the distance conversion accuracy is improved,and the strategy of selecting ranging and location model based on personnel status is realized.Compared with the existing algorithms,the proposed algorithm can adapt to different location environments and has higher accuracy,which realizes the adaptability,universality and high precision of the positioning system.