Research On Indoor Localization Algorithm Based On WiFi Signal

With the development of wireless network technology,location based services have become an irreplaceable part of people's daily lives.Wi Fi localization technology has natural advantages in deployment because it does not need to require additional special equipment.Therefore,this thesis studies the indoor localization algorithm based on Wi Fi signal,which uses the particle filter algorithm and the map topology constraints as auxiliary means to improve the accuracy of localization results.The main work of this subject is as follows:First,this thesis proposes an indoor localization algorithm based on improved particle filtering.The basic localization algorithm in this thesis is the position fingerprint localization method.Different from the traditional fingerprint localization method,this thesis chooses the ray tracing algorithm to construct the indoor signal fingerprint database.At the same time,the self-adaption signal acquisition method is designed to reduce the amount of calculation while ensuring accuracy of the system.Then,aiming at the case that the particle diversity is easy to reduce and even dry up because traditional particle filtering only retains high-weight particles,this thesis designs an auto-optimized resampling method(A-SA)based on the difference in received signal strength at different moments.Finally,in the particle filter fusion phase,the weight of each particle is observed and corrected according to the received signal strength difference between the particles and the target.These particles are weighted to get localization results.The experiments take the resampling method,the total number of APs,and the particle set size as the influencing factors,and prove that the proposed A-SA algorithm can effectively improve the accuracy and robustness of indoor localization without adding any hardware.Secondly,in a narrow indoor environment,the situation where the localization result is in an abnormal position can't be avoided,even if the error is small.Therefore,this thesis proposes a localization algorithm based on map topology constraint(MTC)for indoor environment.The algorithm first divides the traditional indoor error area based on the position of the pedestrian and the APs at the previous moment,and then uses the difference in received signal strength at the target position as its moving state characteristics to build an improved error area model.In order to reduce the situation that correct road section is not found due to small error region,this system uses a rectangular model with road width information to replace the line segment model.Finally,a corresponding candidate link search method is designed based on the improved error area model and rectangular road model.After considering the connection weight,direction weight and distance weight comprehensively,the optimal road segment is determined for correcting the localization result.On this basis,we design and implement an indoor localization prototype system based on Wi Fi signal,which provides a visual interface of the localization effect.In summary,this thesis studies indoor localization algorithm based on Wi Fi signal.Compared with the traditional algorithm,the experimental results prove that the A-SA algorithm improves accuracy by about 35.33%,and the MTC algorithm's mismatch rate decreases by about 9.2%.