Unmanned adjacent vehicle trajectory prediction method

The invention relates to an unmanned adjacent vehicle trajectory prediction method, and relates to the field of unmanned driving. The invention provides an unmanned adjacent vehicle trajectory prediction method. The method comprises the following steps: firstly, extracting a set of vehicles around unmanned adjacent vehicles and road condition information in video data and point cloud data through an LK-DBSCAN (Limit-K DBSCAN) algorithm, constructing potential features with influence by adopting feature engineering, and enhancing the expression ability of the data features to complex road conditions; predicting real-time behaviors of the vehicle by using a long short-term memory (LSTM) neural network; and finally, in combination with the vehicle behavior predicted by the LSTM and the historical behavior data of the vehicle, predicting the trajectory of the vehicle through a B-based LSTM (Behavior-based LSTM). In order to solve the problem, the accuracy of unmanned vehicle behavior prediction can be improved, a