科研成果详情

In the context of edge and fog computing, urban mobility prediction acting as an important role in urban planning, traffic prediction, and resource reservation has making a great contribution for the construction of smart cities, and has been considered to be a challenging research and industrial topic for many years. Generally, those popular prediction methods abstract trajectories into independent points in the manner of gridding, clustering and others. However, these data processing methods make the position representation vector lose the connection relationship between geographic locations which is very important for the mobility prediction. To address this issue, this paper proposes a discrete position relationship model to represent the connection between geographic locations, on the basis, a Long Short Term Memory (LSTM) prediction model is established to predict the next position of the mobile target. Experiments and numerical analyses show that these investigations can take full advantage of the relationship between relative positions and reduce the prediction relative error.