Real-time trajectory prediction method based on Markov process

The invention discloses a real-time trajectory prediction method based on a Markov process. The method is divided into two stages: a data collection stage and a trajectory prediction stage. In the data collection stage, rules of daily work and rest and travel habits of users are extracted from collected historical trajectory data of the users within a long time interval, and the rules are formally constructed into a mathematical model based on the Markov process. In the trajectory prediction stage, the method predicts a future user trajectory in real time based on a mathematical model pre-established in the data collection stage and a user real-time trajectory. The method provided by the invention can accurately predict the user trajectory, can be realized as a sub-module of systems for edge calculation, intelligent transportation, position recommendation and the like, and has wide application value. 本发明公开了一种基于马尔可夫过程的实时轨迹预测方法。该方法分为两个阶段：数据收集阶段和轨迹预测阶段。在数据收集阶段，通过收集的长时间间隔内的用户历史轨迹数据，从中提取出用户的生活作息与出行习惯的规律，并将该规律形式化地构