Compared　with　the　outdoor　vehicle　navigation　system,　the　indoor　environment　is　complex,　the　difficulty　of　positioning　is　greatly　increased.　It　is　effective　to　improve　the　positioning　accuracy　by　combining　the　feature　points　of　the　indoor　building　plane　information　and　the　feature　points　in　pedestrian　walk.　This　paper　chooses　the　intersection　of　the　paths　and　the　turning　points　of　the　pedestrian　movement.　The　intersection　of　the　paths　can　be　given　by　the　building　information　map.　The　turning　points　of　pedestrian　can　be　detected　by　the　difference　of　the　mean　value　of　the　azimuth　(DMA).　If　the　turning　points　are　in　the　corridor,　they　can　be　matched　to　feature　points　and　a　virtual　path　algorithm　is　designed　to　match　the　trajectories.　If　the　turning　points　are　in　the　room,　nothing　is　done　for　them.　Thus　a　new　indoor　pedestrian　location　algorithm　is　obtained.　This　algorithm　can　effectively　distinguish　whether　pedestrian　is　in　the　corridor　or　in　the　room.　Combined　with　the　virtual　path,　pedestrians　in　the　indoor　walking　trajectory　can　be　clearly　demonstrated,　which　decreases　the　positioning　error　of　strapdown　inertial　navigation　and　improves　the　user　experience　of　indoor　positioning　navigation.　©　2017　IEEE.