Acoustic Indoor Localization For Smart Phone With Fusion Of PDR And Map

In recent years,with the rapid development of network communication technologies and intelligent mobile terminals,location-based services have been more and more widely applied to scenes of daily life.In the application field of consumer indoor positioning technology,the positioning technology based on smart phones has incomparable advantages in universality and convenience.In contrast,however,there are still various bottlenecks of high-precision positioning technologies for complex indoor scenes.The positioning technology based on PDR algorithm is not likely to be disturbed by the environment and does not need to install additional base station,but there is cumulative error in the long time positioning.The positioning technology based on acoustic signal can achieve higher positioning accuracy,but the positioning performance will decrease sharply in the non-line-of-sight environment.In short,there is no perfect high-precision indoor positioning system on the market.Therefore,there are still extensive prospects of the research and application on high-precision positioning and tracking under complex indoor scenes.This thesis analyzes the characteristics of common indoor positioning technology from the perspectives of positioning accuracy,robustness,construction cost,transmission distance and mobile phone compatibility.This thesis proposes to use multi-source information fusion positioning algorithm in order to make up for the defects of various single positioning algorithms.Further researches are developed under the framework of the fusion positioning algorithm.The main research contents and innovations of this thesis are as follows.Firstly,this thesis creatively proposes a fusion positioning algorithm based on sound signal measurement,particle filter framework,PDR model and indoor map information.Evaluated by positioning accuracy,construction cost,system robustness and versatility,it appears that the positioning algorithm is superior to other common indoor positioning algorithms.Aiming at the acoustic signal measurement in this algorithm,the performance of each algorithm is verified by simulation,and the position estimation algorithm based on bias reduction for TDOA is selectedSecondly,for the step model of PDR algorithm in the fusion positioning algorithm,this thesis proposes a particle step update algorithm based on the least squares algorithm and the iteration of the positioning result.The step update algorithm does not need to collect user information and it is based on user motion,which make the algorithm easier to apply and more accurate than other step update algorithms.Combined with the TDOA location estimation algorithm based on bias reduction,a PDR-TDOA fusion location algorithm is proposed.Thirdly,for the indoor positioning scene with known map information,this thesis introduces the map information as a constraint condition into the PDR-TDOA fusion positioning algorithm.The particle weight is updated based on the map information,and the target path are constrained with the map information,which has made the positioning accuracy and robustness an obvious improvement.Forthly,for the non-line-of-sight scene of indoor positioning,this thesis proposes a non-line-of-sight recognition method based on acoustic channel propagation characteristics and unsupervised learning.For the misjudgment phenomenon in the method above,in order to improve TDOA location estimation algorithm,this thesis proposes a fusion positioning algorithm based on non-line-of-sight recognition and measurement consistency,which can help to effectively find a reliable reference node and filtrate line-of-sight measurement information.For the scenario of low node coverage,this thesis further proposes a target location estimation method based on non-line-of-sight recognition and map consistency to improve TDOA location estimation algorithm,and uses the improved TDOA position estimation algorithm for fusion positioning.This method makes full use of map information and can be found the best position estimation result with the constraint of map information in the scenario of low node coverage.Finally,based on the positioning algorithm designed in this thesis,this thesis also designs a positioning system which can be applied to the real scene.The positioning system is based on the beacon node and the server arranged in advance,combined with the smart phone of the user end.With this system,real-time indoor positioning result display on the server as well as on the smart phone.