Systems and methods for real-time tracking of trajectories using motion sensors

A system, comprising:

• a memory storing instructions;
• one or more communication interfaces;
• and
• one or more hardware processors coupled to the memory via the one or more communication interfaces, wherein the one or more hardware processors are configured by the instructions to:
• acquire a first sensor data and a second sensor data from a first sensor and a second sensor respectively, wherein the first sensor data and the second sensor data are associated with a user;
• convert the first sensor data and the second sensor data into a first set of scaled sensor data and a second set of scaled sensor data;
• calculate a delta value in a plurality of axis for each sensor value from the second sensor data based on a duration between a current sensor value and a previous sensor value;
• generate a Euler Rodriguez matrix based on the calculated delta value in the plurality of axis for each sensor value from the second sensor data;
• generate a plurality of feature vectors based on the first set of scaled sensor data, the second set of scaled sensor data, and the Euler Rodriguez matrix;
• apply a windowing technique comprising a pre-defined window size on each of the plurality of feature vectors to obtain a plurality of windows, wherein each window amongst the plurality of windows comprises one or more associated feature vectors amongst the plurality of feature vectors;
• and
• train one or more machine learning (ML) models based on a label annotated to each window amongst the plurality of windows, wherein the label indicates a stroke direction of data comprised in each window.

One or more non-transitory computer-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause:

• acquiring, a first sensor data and a second sensor data from a first sensor and a second sensor respectively, wherein the first sensor data and the second sensor data are associated with a user;
• converting the first sensor data and the second sensor data into a first set of scaled sensor data and a second set of scaled sensor data;
• calculating a delta value in a plurality of axis for each sensor value from the second sensor data based on a duration between a current sensor value and a previous sensor value;
• generating a Euler Rodriguez matrix based on the calculated delta value in the plurality of axis for each sensor value from the second sensor data;
• generating a plurality of feature vectors based on the first set of scaled sensor data, the second set of scaled sensor data, and the Euler Rodriguez matrix;
• applying a windowing technique comprising a pre-defined window size on each of the plurality of feature vectors to obtain a plurality of windows, wherein each window amongst the plurality of windows comprises one or more associated feature vectors amongst the plurality of feature vectors;
• and training one or more machine learning (ML) models based on a label annotated to each window amongst the plurality of windows, wherein the label indicates a stroke direction of data comprised in each window.