Mobile device for distinguishing user's movement, method therefor, and method for generating hierarchical tree model therefor

A method for generating a hierarchical tree model for distinguishing a movement behavior of a user having a mobile device, comprising:

• collecting acceleration data separately for each movement behavior from an acceleration sensor within the mobile device;
• dividing the acceleration data for each movement behavior based on a preset time unit to construct a first frame group, and dividing the acceleration data for each movement behavior based on a time unit different from the preset time unit to construct a second frame group, wherein part of the acceleration data of the second frame group overlaps with the first frame group;
• extracting a characteristic factor for each movement behavior that is distinguished from characteristic factors for other movement behaviors, based on the first and second frame groups;
• and
• constructing the hierarchical tree model based on the characteristic factor for each movement behavior.

A method for distinguishing a movement behavior of a user by using a mobile device, comprising:

• collecting acceleration data output from an acceleration sensor within the mobile device according to a specific movement behavior of the user;
• extracting a characteristic factor for the specific movement behavior of the user based on the acceleration data;
• and
• distinguishing to which movement behavior the specific movement behavior of the user corresponds, by inputting the extracted characteristic factor into a pre-constructed hierarchical tree model,
• wherein the hierarchical tree model is pre-constructed based on a characteristic factor extracted for each movement behavior, and
• the characteristic factor for each movement behavior is extracted based on a first frame group and a second frame group, and the first frame group consists of acceleration data collected separately for each movement behavior and divided based on a preset time unit, and the second frame group consists acceleration data collected separately for each movement behavior and divided based on a time unit different from the preset time unit, and wherein part of the acceleration data of the second frame group overlaps with the first frame group.