US 12,452,828 B2
Indoor place prediction
Manoop Talasila, Branchburg, NJ (US); Qiong Wu, Bridgewater, NJ (US); Wen-Ling Hsu, Bridgewater, NJ (US); Xiaopeng Jiang, Kearny, NJ (US); Cristian Borcea, Hillsborough, NJ (US); and Pritam Sen, Harrison, NJ (US)
Assigned to AT&T Intellectual Property I, L.P., Atlanta, GA (US); and New Jersey Institute of Technology, Newark, NJ (US)
Filed by AT&T Intellectual Property I, L.P., Atlanta, NJ (US); and New Jersey Institute of Technology, Newark, NJ (US)
Filed on Oct. 12, 2022, as Appl. No. 18/046,129.
Prior Publication US 2024/0129889 A1, Apr. 18, 2024
Int. Cl. H04W 64/00 (2009.01); G01C 21/20 (2006.01)
CPC H04W 64/006 (2013.01) [G01C 21/206 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
collecting, by a processing system including at least one processor, orientation data from a mobile device as a user of the mobile device moves within an indoor space;
generating, by the processing system, orientation estimates for the mobile device from the orientation data;
generating, by the processing system, distance estimates representing distances of the mobile device to a wireless access point in the indoor space, simultaneously with the generating the orientation estimates;
storing, by the processing system, the orientation estimates and the distance estimates together as a series of data blocks;
constructing, by the processing system, a plurality of trajectory segments from the orientation estimates;
identifying, by the processing system, identifiers for a last number of trajectory segments of the plurality of trajectory segments, using a subset of the series of data blocks corresponding to the last number of trajectory segments, wherein the identifying the identifiers comprises inputting data blocks of the series of data blocks corresponding to the last number of trajectory segments to a classifier that is trained to output the identifiers based on the data blocks that are inputted; and
identifying, by the processing system, a predicted place within the indoor space to which the user is expected to go by using the identifiers to traverse a prediction tree, where branches of the prediction tree that are associated with the identifiers define a path that ends at a leaf node of the prediction tree that corresponds to the predicted place.