Track positioning method for mobile phone positioning deviation correction based on machine learning

The invention discloses a track positioning method for mobile phone positioning deviation correction based on machine learning. The track positioning method comprises the following steps: step 1, determining a physical position; 2, calculating the indoor moving direction and speed by taking an entrance as a track starting point so as to calculate an action track; 3, calculating space floors in the building by using data of the altimeter; 4, accumulating the accurate walking speed of the user through a machine learning algorithm; and 5, accurately performing indoor positioning, correcting the interior by integrating indoor LTE base station signals and positions, calculating the step information and direction of the user through a speed model decision, and then accumulating the accurate walking speed of the user based on the data rate of a machine learning model. And finally, indoor positioning can be accurately carried out through walking speed, orientation, inclination angle and height sensors, and indoor LTE