Design And Implementation Of User Behavior Recognition Subsystem In Full Time-space Fusion Positioning Cloud Platform

At present,due to the popularity of Global Navigation Satellite System,outdoor positioning technology has matured,but for indoor environment,because GPS(Global Positioning System)signal is weak,indoor positioning technology is different from traditional outdoor positioning technology,current indoor positioning technologies mainly include Bluetooth,Wi-Fi,infrared,etc.However,in practice,due to the singularity of these technologies,it is impossible to achieve accurate positioning in some complex scenarios.Therefore,it has been proposed to use the"cloud'*"to integrate various single positioning technologies and to uniformly deploy them through the network.The full-time spatial integration positioning cloud platform is built in this context.The full time-space fusion positioning cloud platfonm has been greatly improved in terns of ubiquity,and can provide continuous,reliable and stable application services in complex scenarios.However,in the actual application process,the following problems need to be solved due to the changing environment in the room.Firstly,the positioning error caused by the change of the user's posture continuously accumulates with the increase of the user's walking time,and positioning effects become worse.Secondly,modern commercial buildings have many complex and irregular internal structures.Therefore,Floor determination required as an auxiliary reduced space search domain.However,the exact switching of floors at the floor decision depends on the current behavior of the user.In order to solve the above two problems,it is necessary to identify the movement behavior of the current user.This paper proposes to add the user's motion behavior recognition subsystem to the full time-space fusion positioning cloud platform.The first objective is to further improve the positioning accuracy by determining the user's current behavior to assist the position calibration in the case of rapid change of the terminal attitude.The other aim is to achieve accurate floor switching by judging the user's behavior,such as go up and down the building,take the elevator,take the escalator,etc.This paper first introduces the background of the user's motion behavior recognition subsystem construction,investigates the relevant implementation technology,and then comprehensively analyze the requirements of the system from both functional and non-functional aspects.On this basis,the overall architecture of the system is given,including functional architecture,physical architecture and technical architecture.And the outline design of function module Division,database and communication interface is carried out.At the same time,the detailed design and coding implementation of the system were completed.Finally,the functional and non-functional testing of the system was completed.This paper successfully completed the entire construction of the user's motion behavior recognition subsystem.As a subsystem of the full time-space fusion positioning cloud platform,the real-time accurate recognition of six movement behavior modes,such as walk on the flat and go up and down the building,improves the positioning accuracy of the fusion positioning cloud platform and greatly promotes its positioning effect in complex scenes.