Abstract:
PURPOSE: A particle filter based user position tracking apparatus and a record medium thereof are provided to track a position of a user fast by re-sampling multiple particles corresponding to the candidate position of a user. CONSTITUTION: A user position tracking part includes a particle initializing part(200), a particle moving part(202), a likelihood calculating part(204), a weighting factor installing part(206), and a particle re-sampling part(208). In case a user moves in an inner space, a particle moving part moves each of multiple particles from a previous position to a new position according to a predetermined movement model. A likelihood calculating part calculates the likelihood of multiple particles by using wireless signal information received from one or more wireless AP in an actual position which a user moves and wireless signal map produced in advance. A weighting factor installing part renews each weighting factor of multiple particles by using a calculated likelihood. A particle re-sampling part selects a particle to replace multiple particles by using a renewed weighting factor. [Reference numerals] (200) Particle initializing part; (202) Particle moving part; (204) Likelihood calculating part; (206) Weighting factor installing part; (208) Particle resampling part; (AA) User position tracking unit
Abstract:
PURPOSE: A system and a method for predicting a route of a user based on a GPS log and a learned route model are provided to easily predict a route where a user is currently moving on based on a current GPS log of the user and a study route model learned by the user in advance. CONSTITUTION: A method for predicting a route of a user based on a GPS log and a learned route model is as follows. A candidate learning route within a minimum boundary range from a current moving coordinate set of a user is extracted among learning routes of the user of a learning route DB(S110). The similarity between the current moving coordinate set and the candidate learning route is calculate so that the candidate learning route in which the similarity is the highest is decided(S120). The validity with respect to only the decided candidate learning route(S130).