Multisource Fusion Indoor Positioning Algorithm Based On Adaptive Particle Filter

With the rapid development of information technology,location information has become important information in modern people's daily life,and location-based services have been widely used in commercial promotion,smart city and other fields.Since people spend most of their time in indoor environment,indoor location information can effectively reflect personal habits and consumption tendencies,and has great application prospects in building user portraits,guiding user's consumption and improving user 's experience,thus indoor positioning technology has become a hot research topic in recent years.A detailed analysis and review of indoor positioning technologies is presented from three aspects: indoor positioning sources,single-source indoor positioning methods,and multisource fusion indoor positioning methods,including signal characteristics of indoor positioning sources,hardware overhead,indoor positioning accuracy,advantages and disadvantages of being an indoor positioning signal source,and key attributes are summarized.Single-source indoor positioning methods are classified,the positioning principles and algorithms are analyzed,and the advantages and disadvantages of single-source positioning methods are discussed;multi-source fusion indoor positioning methods are studied,and the advantages and disadvantages of multi-source fusion positioning technologies are analyzed and compared.To solve the problems of high computation cost and high fusion delay caused by a large number of sample particles to approximate the probability density function in multisource fusion indoor positioning algorithm,an adaptive particle filter based multisource fusion indoor positioning algorithm(APFP)is proposed.Considering the system cost and equipment support issues,the Bluetooth beacon and the built-in sensor of the mobile phone are adopted as the positioning signal source,and Bluetooth RSS(Received Signal Strength)trilateral positioning and cell phone PDR(Pedestrian Dead Reckoning)positioning are selected as the single-source positioning methods.The adaptive filtering conditions are designed by combining the localization characteristics of two single-source localization methods.APFP effectively combines the characteristics of RSS trilateral positioning error free accumulation of Bluetooth and PDR positioning short-range accuracy of mobile phone.It only performs filters when the conditions of adaptive filtering are satisfied,which reduces the filtering times,and the fusion delay on the premise of ensuring the fusion accuracy.The performance of Bluetooth RSS trilateral positioning and mobile phone PDR positioning,as well as traditional particle filtering algorithms,particle filter algorithm based on map constraints,and the proposed APFP is evaluated with MATLAB,in terms of RMSE(Root Mean Squared Error),MAE(Mean Absolute Error),and time delay.A Bluetooth beacon positioning testbed was setup,and experimental data was collected in the testbed.The experimental results shows that,compared with the traditional particle filtering algorithm and the particle filtering algorithm based on map constraints,in the low-noise indoor environment,APFP reduces the fusion delay by 59.89% and54.37%,respectively,while guaranteeing the localization accuracy;in the high-noise indoor environment,compared with the traditional particle filtering,APFP not only has higher localization accuracy,but also reduces the fusion delay by 53.1%.